* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download PDF - Routledge Handbooks Online
Arabic grammar wikipedia , lookup
Malay grammar wikipedia , lookup
Modern Greek grammar wikipedia , lookup
Old Norse morphology wikipedia , lookup
Swedish grammar wikipedia , lookup
Ancient Greek grammar wikipedia , lookup
Scottish Gaelic grammar wikipedia , lookup
Yiddish grammar wikipedia , lookup
Old Irish grammar wikipedia , lookup
Old English grammar wikipedia , lookup
Icelandic grammar wikipedia , lookup
Serbo-Croatian grammar wikipedia , lookup
Portuguese grammar wikipedia , lookup
Polish grammar wikipedia , lookup
French grammar wikipedia , lookup
Italian grammar wikipedia , lookup
Spanish pronouns wikipedia , lookup
Spanish verbs wikipedia , lookup
Latin syntax wikipedia , lookup
This article was downloaded by: 192.168.0.15 On: 12 Jun 2017 Access details: subscription number Publisher:Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place, London SW1P 1WG, UK The Romance Languages Martin Harris, Nigel Vincent Spanish Publication details https://www.routledgehandbooks.com/doi/10.4324/9780203426531.ch3 John N. Green Published online on: 28 Jan 1988 How to cite :- John N. Green. 28 Jan 1988 ,Spanish from: The Romance Languages Routledge. Accessed on: 12 Jun 2017 https://www.routledgehandbooks.com/doi/10.4324/9780203426531.ch3 PLEASE SCROLL DOWN FOR DOCUMENT Full terms and conditions of use: https://www.routledgehandbooks.com/legal-notices/terms. This Document PDF may be used for research, teaching and private study purposes. Any substantial or systematic reproductions, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The publisher shall not be liable for an loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 3 Spanish John N. Green 1 Introduction Spanish is by far the most widely spoken of the modern Romance languages, and as an international vehicle for commerce and diplomacy is fast encroaching on the preeminent position long enjoyed by French. (Details of its geographical distribution together with speaker statistics will be found in Chapter 1. See also Map I.) In most Spanish-speaking regions the terms espanol and castellano are used interchangeably, but a useful distinction can be drawn between 'Spanish' as a diasystem and 'Castilian' as its prestige form in Europe, the basis of the standard language and still, for many speakers, the model for pronunciation. In common with all spatially diffused languages, Spanish is subject to regional and sociolinguistic variation and also — in its international role — to conflicting normative pressures, but despite some well-publicised divergences of pronunciation and vocabulary (discussed in sections 2 and 5 below), the range of variation is not very great and only rarely disrupts mutual comprehensibility. In Spain, the 'purest' form of Castilian is traditionally identified with Burgos, but in practice the norm has long been the educated usage of Madrid, which has consistently been more open to outside influences. In Latin America, the prestige formerly attaching to Colombian Spanish (perhaps because Colombia was the first ex-colony to establish its own language Academy in 1873) has now unquestionably moved northwards to Mexico City, by far the most populous Spanishspeaking conurbation in the world. The modern focus of the two 'norms' does not, however, mean that their geographical boundary can be neatly located between Europe and the New World, since Andalusian in southern Spain and the dialects of the Canary Islands share many of the distinguishing features of Latin American. Despite the vast discrepancy in populations, the Castilian norm continues to be highly valued — though not necessarily imitated — in most of Latin America. An ironical recent development in Spain, however, has been the spread, among younger speakers, of modes of pronunciation whose origin is unmistakably Andalusian. 79 Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 80 SPANISH 2 Phonology In this sketch of Spanish phonology we shall concentrate on the Castilian norm, making reference as appropriate to variant articulations and sociolinguistic status. An arrangement of material adopted in the interests of economy should not, however, be taken to imply that other varieties of Spanish are deviations from, nor yet dialects of, Castilian. Auditory Impression Before considering the segmental units, we should say a little about prosodic characteristics and general auditory impression. Spanish has often been quoted as a textbook example of a syllable-timed language, with a delivery sometimes likened to a recalcitrant machine gun. A newer proposal suggests Spanish would be more accurately described as 'segment timed' since the delivery, though perceptually regular, does not always produce isochronous syllabification or isochronous stress intervals. The rhythmic pattern, naturally, has implications for intonation, which tends to avoid abrupt changes and readily accommodates melodic units of ten to fifteen syllables. Castilian, whose everyday register is confined to little more than an octave, has a basic rise-fall for simple declaratives, a sustained rise for most yes-no questions, and the characteristic Western Romance level or rising tone to mark enumerations and sentence-medial clause boundaries. A prominent feature of Castilian is its 'dynamic' or intensity accent, which is noticeably free from tonal modulation. Most writers also comment on the resonant quality that Castilians and northern dialect speakers impart to their everyday speech. This has been variously ascribed to an unusual articulatory setting, to the rhythmic structure, to the predominance of low, open vowels, and to the stability of vowel sounds in both stressed and unstressed positions. Though all these factors may be contributory, the principal cause must be articulatory setting, since many other regional varieties of Spanish are produced with a less marked resonant quality despite sharing the other structural features of Castilian. Consonants The segmental consonant system of Castilian, given in Table 3.1, can be presented as neatly symmetrical, with four articulatory positions and five degrees of aperture, but this disguises some interesting irregularities in distribution. While, for instance, the absence of any point-of-articulation opposition between oral stops and affricates argues for a merger of the categories, they differ in that /dz/ is by no means securely established in the system, and neither palatal enters into syllable-initial clusters, which the stops do freely. The reintroduction of [dz], which was present in Old Spanish probably as an allophone of /z/, is comparatively recent and its phonemic status remains doubtful. Since, as we shall see (p. 84), it occurs as the exponent of two distinct phonemes, its incorporation would neces- Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 81 Table 3.1: The Castilian Consonant System Oral stops [—voice] [+voice] Affricates [—voice] [+voice] Fricatives [—voice] Nasal stops Laterals Vibrants [+tense] [—tense] Bilabial Dento-alveolar P t b Palatal k d tf f m θ Velar s n 1 g (dz) X η λ r r sarily provoke a realignment of the system. The Castilian voiceless stops are unaspirated plosives. The voiced series is in complementary distribution with a corresponding set of voiced spirants, the plosives occurring word-initially and medially after nasals, and the spirants elsewhere: boca 'mouth' ['boka] but cabo 'end' ['kaβo], donde 'where' ['donde] but nudo 'knot' ['nudo], grande 'big' ['grande] but magro 'lean' ['mayro]. A word-initial plosive, however, is also liable to replacement by the corresponding spirant when intervocalic within a breathgroup: la boca [la'βoka]. Moreover, in indigenous words neither / b / nor / g / occurs word-finally, and orthographic -d is weakened to [d] or lost completely. It has traditionally been assumed that the spirants are the subordinate members of these pairs, since the weakening of plosives to fricatives in comparable environments is well attested as a historical process in Romance. But the distinction is equally amenable to analysis as a strengthening of spirants in group-initial or post-nasal position, and recent research on language acquisition among Mexican children seems to show that the spirants are acquired first and remain dominant. The future of the medial spirants is, however, less secure than their present frequency implies. The Latin intervocalic voiceless plosives from which many of them derive have in French often undergone a third stage of lenition resulting in their complete loss and a consequent syllabic merger (see p. 39), compare VĪTAM 'life' > Sp. ['bida], Fr. [vi]; MĪCAM 'crumb' > Sp. ['miya], Fr. [mi]. There are a number of pointers to a similar outcome in Spanish. One is the very lax articulation often given to intervocalic / - g - / especially when followed by the labio-velar glide / w / ; in a word like agua 'water' the g is rarely more than a frictionless velar approximant and Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 82 SPANISH is sometimes lost altogether. Another sign is the disappearance of /-d-/ from past participles in -ado/-ido. This change, which is now virtually complete among younger speakers, has been the subject of intense normative disapproval, giving rise in turn to amusing hypercorrections like [baka'lado] for bacalao 'cod'. Curiously, the same change has gone unnoticed in the reflexive imperative, where the coalescence of levantad + os as ilevantaos! 'stand up' (familiar plural) is even accepted by the standard orthography. The loss of /-d-/ in these contexts is still clearly dependent on morphological factors (in the past participle, it affects the feminine form -ada less frequently than the corresponding masculine -ado), but it may signal the first stage of a more general phonological process. The three Castilian sibilants are the remainder of what was in Old Spanish a much larger set, including a voiced phonemic series whose demise is still not wholly explained. The absence of phonemic voiced sibilants now sets Spanish apart from most other Romance varieties, though [z] is used by some speakers as an alternative realisation of /λ/ (see below), and [z] occurs infrequently as an allophone of / s / before voiced obstruents, but not intervocalically, thus desde 'from/since' ['dez e] but esposa 'wife' [εs'posa] — compare Port. ['ƒpoza], It. ['spc:za] and Fr. [e'puz]. In many non-Castilian varieties, the inventory of sibilants, or their frequency, is further reduced, for example, by the aspiration and sometimes complete loss of syllable-final / - s / (see p. 85). Moreover, American Spanish and most varieties of Andalusian lack the distinctive Castilian opposition between / θ / and / s / , as in cima 'summit' /θima/ : sima 'abyss' /sima/, caza 'hunt' /kaθa/ : casa 'house' /kasa/, haz 'bundle' / a θ / : as 'ace' /as/. Throughout South America and in most of Andalusia, only [s] is found — a feature popularly called seseo. The converse, using only [θ] and duly known as ceceo, is confined to scattered locations in Andalusia. The assumption has been quite widely made that the majority seseante territories lost an earlier opposition between / θ / and / s / and therefore represent innovatory tendencies in relation to the conservatism of Castilian. (It has indeed been suggested that American Spanish still maintains an opposition at underlying level which is always neutralised in phonetic representation — an idea further discussed on p. 101 below.) The truth is certainly more complex. From the available documentary evidence it cannot be conclusively established whether or not a phonological opposition between two front sibilants was present in the Spanish of the first colonists, but it is now almost certain that the interdental fricative articulation [8] was a Castilian innovation which significantly post-dated the early phase of settlement. Seseante regions may, therefore, have eliminated something since the sixteenth century, but not specifically the [θ] sound. How then did it arise in Castilian? The most plausible explanation is: as a way of increasing the auditory distinctiveness of an opposition which was in Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 83 danger of coalescing. The loss of distinctive voicing had reduced the six medieval sibilants to three by the fifteenth century; let us call them S1, S2 and S3. Suppose now that S, had been simplified from a medieval affricate /ts/ to a dorso-alveolar or laminal / s / ; that S2 represented a less common apico-alveolar / s / ; and that S3 was a palatal /ƒ/. The distance — both articulatory and auditory — between the three sounds would have been very small. In the circumstances a natural development would be the 'splaying out' of tongue positions into: front apical [θ], mid apical [s] and back dorsal [x]. These values are very similar to the ones now found in modern Castilian, except that the /x/ phoneme is often realised even further back as a uvular [χ], probably to maximise the difference between /x/ and / s / , whose unusual apico-alveolar articulation gives a slight auditory impression of palatality. By contrast, Mexican Spanish has a noticeably forward articulation of velar /x/, which is in no danger of coalescing with the only other sibilant, a forward laminal / s / whose distribution echoes that of the former S, and S2. If the reconstruction we have just outlined is correct, we shall have to say that seseante dialects are innovatory in relation to late medieval Spanish (having eventually reduced six sibilants to two), whereas Castilian is conservative in its phonology (having maintained a three-way opposition) but still innovatory in its phonetics! The merger of S, and S2 in seseante varieties, whatever its exact date, naturally created homophones, some of which have persisted. But in other cases it seems to have led to lexical substitutions in order to avoid ambiguity: the Castilian minimal pair coser 'to sew' /koser/ : cocer 'to cook' /koθer/ poses no problem in America, where coser is maintained, but cocinar /kosinar/ is the verb 'to cook'. Let us now turn to the Castilian sonorants. The three nasals contrast intervocalically, where there are numerous permutations of minimal pairs and a few triads: lama 'slime' /lama/ : lana 'wool' /lana/ : lana 'clamp' /laηa/. Elsewhere, the opposition is incomplete. Word-initially, / η - / is very rare, confined to a few affective terms and borrowings from Amerindian languages; among Latinate items, only /m-/ and /n-/ are possible. Nasals combine freely with obstruents to form heterosyllabic clusters, in which seven or more phonetic variants can be detected, always homorganic with the following consonant and therefore neutralising the opposition — infeliz 'unhappy' [imfe'liθ], incierto 'uncertain' [in1θεrto], incapaz 'unable' [iηka'paθ], and so on. The opposition is also neutralised in word final position, where only /-n/ occurs. A variant pronunciation, previously common in Andalusia and parts of Latin America, is now spreading rapidly in Spain, though it remains sociolinguistically marked: wordfinal and sometimes syllable-final /-n/ is realised as velar [-η], with appreciable nasalisation of the preceding vowel. (For similar developments in Italian, see p. 282.) It has been argued that the weak articulation of final Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 84 SPANISH [-η] may be the prelude to phonemic nasalisation of the vowel, a development Spanish has so far resisted. Table 3.1 shows an opposition between two lateral phonemes, a dentoalveolar / l / and a palatal /λ/. Though it does not carry a high functional load, the opposition is regularly maintained in such pairs as low 'parrot’ : lloro 'I weep'. The phonetic realisation of /λ/, however, is subject to wide variation, and for many speakers the pronunciation [λ] is now characteristic only of very formal or careful delivery. The delateralisation of /λ/ to [j] as in French (p. 212), long familiar in Latin America under the name yeismo, and now spreading rapidly in Spain, has created an allophonic overlap with the glide / j / , and consequently a new set of homophones — like ['pojo] for both polio 'chicken' and poyo 'bench' — affording rich possibilities for confusion and hypercorrection. A secondary development of yeismo strengthens the [j] derived from /λ/ to [z] or even [dz], but in Spain these further stages remain sociolinguistically marked: whereas in Madrid the pronunciation of words like calle 'street' as ['kaje] is now virtually standard, the fricative variant ['kaze] is often regarded as uneducated and the affricate realisation ['kadze] is usually stigmatised as vulgar. Less stigma, however, seems to attach to an earlier, and apparently independent, strengthening of certain initial [j] sounds to [dz], a prominent example being the pronunciation of the personal pronoun yo‘I’as [dzo] rather than [jo]. The long-term effects of this series of changes are by no means clear. The generalisation we can properly draw in the interim, is that only syllable-initial [j], whatever its source, is a possible input to strengthening. For as long as this holds good, there will not be a complete coalescence of /j/ and /λ/. The vibrants / r / and / r / are the only sonorants to contrast at the same point of articulation, but the opposition is only intervocalic —caro'expensive' /karo/ : carw 'cart' /karo/; elsewhere the two sounds are in complementary distribution. In standard Castilian the difference seems to be one of tenseness rather than length: /-r-/ is usually a flap and /-r-l a fullbodied aveolar trill. The /r:r/ opposition is maintained in all dialects, but with a very wide range of phonetic exponents. Particular mention should be made of the fricativisation of / r / to velar [x] or uvular [R] in Puerto Rico and Colombian coast Spanish, together with its realisation in some other American varieties as a weak trill with palatal friction [r]. This latter development (also observed in Brazilian Portuguese, p. 138) suggests an intriguing historical parallel with the palatals / η / and /λ/, since all three derive principally from Latin intervocalic geminates and appear to have evolved via a stage of tenseness. A second parallel could be drawn between /λ/ and /r/ in that their most recent (and still sociolinguistically marked) evolution seems likely to remove them from the class of sonorants. Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 85 Vowels As is apparent from Table 3.2, the five simple vowels of modern Spanish form a classic symmetrical triangle. Their frequency of occurrence in running prose also follows a regular pattern: low vowels are more frequent than high, front more so than back (hence in ascending order /u, i, o, e, a/). All five occur as independent words, with / e / and / a / both representing homophones. All occur both stressed and unstressed, in open and closed syllables, though / i / and / u / are rare in word-final position. As we noted above, there is little tendency to weakening or centralisation in unstressed syllables, a feature which sets Spanish clearly apart from its peninsular neighbours Portuguese and Catalan. Regardless of the presence or absence of stress, however, all vowels are represented by laxer variants in closed syllables; the high and mid series are lowered slightly and / a / , which in citation has a central low articulation, may be displaced forward or backward depending on the adjacent consonant: preste 'I lent' /pres'te/ [prεs'te], corto 'it cut' /kor'to/ [kor'to], jaulas 'cages' /'xawlas/ ['χawlaes]. Table 3.2: Vowels and Semi-vocalic Glides High Mid Low Vowels Glides i e j a u o w This unexceptional laxing has paved the way for a change in Andalusian and some Latin American varieties which may have far-reaching consequences for the vowel system and for plural marking. The great majority of Spanish nouns in the singular end in open /-a/, / - o / or / - e / , but the addition of the plural marker / - s / closes the syllable and produces the regular allophonic variation in the vowel: hermano(s) 'brother(s)' /ermano/ [εr'mano] + /s/ — [εr'marics] hermana(s) 'sister(s)' /ermana/ [εr'mana] + /s/ = [εr'manaes] madre(s) 4mother(s)' /madre/ ['ma re] + /s/ — ['ma res] In Andalusian, syllable-final / - s / often weakens to an aspirate [-h], so los hermanos becomes [loh εr'manch] and so on. This substitution, though phonetically salient, does not affect the phonemic status of the vowels. In a more 'advanced' variety of Andalusian, however, the aspiration is lost altogether and with it the conditioning factor for the vowel alternation. Now [la 'ma re] contrasts functionally with [lae 'ma rε] in a new system of plural marking not too far removed from the vocalic alternations of Italian (see p. 289). The implications for the phonological inventory are disputed: on the Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 86 SPANISH one hand the new functional contrasts argue for the recognition of three new vowel phonemes, but this has been countered by the claim that Andalusian plurals are now marked by a prosody of laxing which affects all vowels in the word, not merely the one in contact with historical / - s / . Table 3.2 above shows no diphthongs or triphthongs. On the phonetic level, combinations of vowels and vowel-like elements are common, but their phonemic status has always been among the most controversial areas of Spanish linguistics. Eighteen monosyllabic combinations can be distinguished, eight with a glide onset /ja, je, jo, ju, wa, we, wi, wo/, six with an off-glide /aj, aw, ej, ew, oj, ow/, and a further four with both on- and offglides /waj, wej, jaw, waw/ of which the last two are very rare. The analyst's task is complicated by the existence of numerous other combinations, both within and across word boundaries, of vowels 'in hiatus' — pronounced as two syllables in careful speech, but readily coalescing into monosyllables in rapid or informal delivery. To explain the controversy, we must make a brief foray into stress assignment. Stress Assignment Stress in Latin was non-phonemic and predictable from syllable structure (see p. 31). Throughout the history of Spanish, stress has remained extraordinarily faithful to its etymological syllable, though as a result of phonetic evolution that syllable may no longer bear the same relation to the whole word as it did in Latin. A good example is provided by Latin words regularly stressed on the antepenultimate syllable, almost all of which have lost their intertonic vowel and now conform to the majority pattern of penultimate stress: SAL CEM /'sa-li-ke/'willow' > sauce /'sawθe/. Like Latin, modern Spanish does not use stress to differentiate lexical items, but the cumulative effect of phonetic change has been to remove some of the previous conditioning environments, so that stress placement is no longer wholly predictable from surface phonetic structure. The simplest phonological treatment thus entails marking stress as an inherent property of lexical items. Most linguists nevertheless feel this root-and-branch approach denies significant generalisations. In this belief they are supported by the standard orthography, which uses a written accent only when stress deviates from the expected position, defined as follows: words ending in a consonant other than -s or -n are stressed on the final syllable, others (the great majority) on the penultimate. The exemption of -s and -n, which serve as plural markers for nouns and third person verbs respectively, is designed to accommodate the fact that plurals almost invariably maintain the stress pattern of their corresponding vocalic singulars: la(s) chica(s) corre(n) 'the girl(s) run(s)' /la(s) 'tƒika(s) 'kore(n)/. The orthographic rules automatically define as an exception any form stressed earlier than the penultimate syllable, including the not insignificant number of proparoxytones (known in Spanish metrics Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 87 as esdrùjulas) reintroduced into the language by lexical borrowing. In practice, the orthographic conventions work reasonably well, but they do leave a residue of quite common nouns and adjectives which appear to be stressed in the ‘wrong’ place: àrbol‘tree’, cdrcel‘prison’, césped‘turf,’ lapiz 'pencil', mdrtir 'martyr', útil 'useful', all have penultimate stress where final might be expected, and the converse applies in: holgazán ‘lazy’, pais 'country', preguntón 'inquisitive', and so on. Some allegedly phonological analyses clearly derive from these orthographic conventions. The properly phonological generalisations thus captured are: that the preferred stress position in Spanish is penultimate; that stress can move further back than the antepenultimate syllable only if the word is clearly composite, like entregdndomelo 'handing it to me' or fácilmente 'easily' (though such adverbs have a secondary stress in the expected position); that plurals maintain the pattern of singulars even when this entails creating proparoxytones like jdvenes 'youths'; and that otherwise antepenultimate stress is not phonologically predictable (the open penultimate syllable common to all such forms being a necessary, but not sufficient, condition). Even so, it should be clear from the examples already quoted that the seemingly capricious stress-assigning properties of -n, -r and -5, in fact depend on whether these segments form part of the lexical stem, or represent suffixal morphemes. Whatever interpretation is placed on this correlation, the fairly straightforward account of stress we have so far envisaged will necessarily be complicated when we turn to verbal inflection. Here, stress operates functionally to differentiate otherwise identical forms of the same lexeme — hablo ‘I speak’ : hablo '(s)he spoke', icante! 'sing!': cante 'I sang', tomara ‘(s)he would take’: tomard '(s)he will take'. It follows that an analysis wishing to view stress as generally predictable must make reference to morphological information. Some theories, of course, rule this out by axiom. Semi-vocalic Glides The status of the semi-vocalic glides is problematic because, although [j, w] appear to be in complementary distribution with the vowels /i, u/ respectively, this economical analysis requires prior knowledge of stress position: / i / is realised as [j] (or becomes [-syllabic]) if and only if it is unstressed and adjacent to some other vowel. The neatest analysis of all would thus need to show that glide formation and stress assignment are both predictable. Can it be done? Consider these examples: amplio [‘am-pljo] : amplío [am-'pli-o] : amplio [am-'pljo] 'ample' ‘I broaden' '(s)he broadened' continuo [kon-'ti-nwo] : continúo [kon-ti-'nu-o] : continu6 [kDn-ti-‘nwo] ‘continuous' ‘I continue' ‘(s)he continued’ Here, the occurrence of the full vowel or glide is predictable, once stress is Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 88 SPANISH known. But the converse is not true: stress cannot be predicted using only the phonological information given here. Nor can it be made predictable by including general morphological conditions, since other verbs behave dif ferently in the middle form of the series: cambiar ‘to change’ and menguar ‘to lessen’ conjugate respectively as ['kam-bjo] and ['merj-gwo] not *[kam-'bi-o] or *[mεη-,gu-o]. For reductionists, the consequences are uncomfortable: neither glide formation nor stress assignment can be pre dicted on phonological criteria alone. Systematic Alternations An allied debate has raged around the predictability or otherwise of the verb stem alternations traditionally called 'radical changes'. The two most frequent ones involve glides and stress assignment. The verb poder 'to be able' has two stems: /pod-/ when the following vowel is stressed and /pwed-/ when the stem itself is stressed. This results in a heterogeneous paradigm, very striking in the present indicative, with 1 sg. puedo alongside 1 pi. podemos. Similarly, helar 'to freeze', has the stressed stem hielo /jelo/ alongside helamos /elamos/. Some 400 verbs follow these two patterns, far more than one would normally wish to describe as 'irregular'. In any event, the observable changes are perfectly regular once the stress assign ment is known. But the interesting question is whether membership of the radical changing pattern is itself predictable. It used to be. The seven-term vowel system inherited by most West Romance dialects (p. 33) main tained a phonemic contrast between the pairs of mid vowels /e: e/ and / o : o/. In northern Spain, /ε/ and /o/ diphthongised when stressed. This was a regular phonological change, affecting all word classes equally and all types of syllable (in northern French, the same vowels diphthongised only in open syllables; see p. 211). So, Spanish verbs with /ε/ or /o/ as their stem vowel were regularly subject to diphthongisation under stress, that stress in turn being positioned according to the number of syllables in the inflection. What has changed between early and modern Spanish is the loss of the phonemic opposition between the mid vowels in favour of an allophonic variation predictable from syllable structure. It is no longer possible to tell, from an infinitive, whether a verb will be radical changing or not: the stem vowel of podar 'to prune' is identical to that of poder but does not diph thongise; neither does the e of pelar 'to peel', although it is phonetically indistinguishable from that of helar. These facts are susceptible of various competing analyses. Some linguists, arguing that so common an alternation must result from a regular productive rule, have postulated underlying vowels /ε, o/ for radical changing verbs and thus claim the synchronic process is identical to the historical change. The theoretical elegance of this treatment is not in doubt, but it is by no means clear that Spanish language learners could Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 89 reconstruct from imperfect synchronic alternations an underlying opposition which is always neutralised in phonetic representation. Their task, indeed, would be even harder: the opposition they need to reconstruct — and later neutralise — is apparently present at the phonetic level as a result of the laxing process routinely affecting closed syllables (see p. 85). The phonetic input is thus most confusing. Not only do learners have to postulate underlying /o/ in poder but not podar when both are actually pronounced with [o]; they also need to distinguish between torcer 'to twist', which diphthongises, and cortar 'to cut', which does not, even though both have phonetic [D] in the infinitive! Other linguists, opposed in principle to such abstract analyses, have argued that the alternation is real, but works synchronically in the opposite direction. If forms such as hiela or puede are taken to be basic, rather than the traditional infinitive, then the diphthong can be made to 'revert' automatically to a simple vowel whenever stress is displaced to another syllable. For the data so far adduced, this proposal works well, though there would be a handful of exceptions: juega '(s)he plays' alternates with jugar not *jogar, and a verb like licuece 'it liquifies', whose diphthong derives from a different historical source, does not revert to *licocer. The proposal fails, however, to accommodate a less common radical change in which unstressed / e / alternates with stressed / i / , as in pedir 'to request' : pide, medir 'to measure' : mide. If the / i / alternants here are taken as basic, it becomes impossible to explain why vive '(s)he lives' has its infinitive in vivir rather than *vevir. Even if these difficulties could be overcome, the proposal would — like the abstract analysis it aims to supersede — still require access to morphological information. This is because the alternation is much less regular outside the verb system than within it. The adjective bueno 'good' has an associated noun bondad in which the final syllable is stressed and the diphthong duly replaced, but its superlative may be either bonisimo or buenisimo. Other derivational processes are undermining the earlier phonological regularity of diphthongisation: deshuesar 'to remove bones/pits' is a verb coined from the noun hueso, but the diphthong which regularly occurs under stress in the noun is irregular in the infinitive, where it is unstressed. Parallel examples are ahuecar 'to hollow out' from hueco, or amueblado 'furnished' from mueble. Not surprisingly, a third body of opinion believes that Spanish speakers cannot predict these alternations at all, and must learn them as inherent features of the individual word. This view is supported by the observation that speakers of some varieties stigmatised as substandard, notably Chicano, regularly keep the diphthongised stem throughout a paradigm regardless of stress placement, saying despiertamos for standard despertamos 'we wake up' and recuerdamos for standard recordamos 'we remember'. A recent experiment with speakers of standard American Spanish has provided further corroboration. Informants were given a set of connected Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 90 SPANISH sentences containing nonce verb forms and asked to complete other sen tences which required different grammatical forms of the made-up verbs. The informants proved most reluctant to ‘create’ diphthongs or other irre gularities even in the nonce verbs very reminiscent of frequent 'real' verbs; but more surprisingly, many also resisted replacing a given diphthong by its corresponding simple vowel in unstressed positions. All told, it looks as though a process which at first was phonoiogically regular has passed through a stage of morphological conditioning and is now giving way to lexical marking on individual words. As often happens in linguistic change, this will preserve analogical relationships at the expense of phonological regularity. Orthography Spanish orthography, though popularly reputed to be 'phonetic' (by which is meant 'phonemic') is in fact quite highly conventionalised. The corre spondence of letters to sounds is skewed by the preservation, and occasion ally restoration, of etymological spellings. Once the conventions have been mastered, it is relatively easy to read aloud the written language; but tran scribing from speech is altogether trickier, as attested by the difficulty Spanish schoolchildren experience with dictation exercises. Standard orthography is also biased towards the Castilian norm, making its acqui sition more difficult for the great majority of speakers who have no phono logical opposition between /s/ and /θ/ (see p. 82). As can be seen from Figure 3.1, which assumes a conservative articu lation, only six letter-sound correspondences can be considered strictly biunique. Obvious difficulties for transcription are: the invariably mute h, the velar consonants, the alternation of c : z and g : j, and the distribution of b : v. Since b and v correspond to only one phoneme and are not in the same distribution as its two allophones — beber 'to drink' = /beber/ [be'βεr],vivir'to lie' = /bibir/ [bi'βir] — it is hardly surprising that words containing these letters are often misspelled, even on public notices; two recently observed in Segovia province read Se prohive aparcar [= prohibe] 'no parking' and Coto pribado de caza [= privado] 'private hunting'. Among the velars, the orthographic contexts for the two values of c and g recapitulate the phonetic environment for early Romance palatalisation, but preserving this relationship in the modern language results in asym metries. For instance, the reluctance to employ k in native words requires the use of the digraph qu to represent /k/ before front vowels. This pro duces a somewhat spurious parallel in verbs which, though morphologically regular, require a modification in spelling when the vowel of the ending changes, hence: indicar 'to indicate' becomes indique in the subjunctive and otorgar 'to grant' becomes otorgue. But qu, which occurs only before i, e, is a genuine digraph, while gu is bivalent: in gui,gue /gi,ge/ it acts as a digraph, but before non-front vowels it represents the sequence /gw/, Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH Figure 3.1: The Principal Letter-Sound Correspondences of Castilian Sound Phoneme Letter/digraph Orthographic context Allophone Phonological mnte.rt 0 h (h)a (h)e (h)o (h)u a e o u w w (h)i _c,# _v i v -J #y#' 1 11 I r rr r /# elsewhere f n n m +v,f \P n / m P p +sb v f ch s b f # elsewhere -p /# ,/N_ elsewhere •tf s -z t d / c V t d N -3 3 z c qu k J g gu _+i,e +a,o,u,C — +i,e — +i,e- -+a,o,u,C +i,c elsewhere 8 ~k X i. N_ elsewhere -k+s Key: / — in the environment;, = or; # = word boundary; + = followed by; V any vowel; C = any consonant; C — any voiced consonant; N — any nasal consonant. 91 Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 92 SPANISH which in turn requires the sequences /gwi, gwe/ to be spelled with a diaeresis as guit giie. A further asymmetry arises in the verb system as a result of the preference for representing / 6 / as c before front vowels, in which environment z is restricted to a few foreign words such as zinc and zebra. Hence, verbs like alzar 'to raise' are spelled with c in the subjunctive {alce), perhaps to counterbalance verbs like esparcir 'to scatter' which in the corresponding forms have no alternative to using z (esparza). But the parallel does not hold good for g and / as representations of /x/: while acoger 'to welcome' must use /"in the subjunctive (acoja), dejar 'to leave' retains its j(deje), spurning the available alternative of *dege. These and other anomalies are the indirect consequences of the wish to maintain an etymological relationship elsewhere in the system. Sometimes, more than one etymological 'layer' is preserved: h, for instance, represents Latin h in hombre 'man' < HOMINE (variously spelled omne/ omrel ombre in Old Spanish), Latin / i n hablar 'to speak' < FABULARE (see p. 121), Arabic / h / in alhena 'henna' < /al+hinnaa?/, and Arabic /{/ in alhondiga 'granary' < /al+funduq/. Whether current Spanish orthography is the most efficient representation of the spoken language remains a matter of debate, since the answers vary according to desiderata. If the prime aim is to preserve the world-wide visual uniformity of Spanish, some departure from phonemic principles is unavoidable. One should not underestimate the impact of a spelling reform, particularly one not shared by all Spanishspeaking countries: at first sight Djudeo-espanyol, in its near-phonemic Israeli orthography, is a different language. We should also note that if etymological relics were eliminated from all Romance languages, their visual relatedness would disappear overnight. 3 Morphology It is difficult in many languages to draw a principled distinction between morphology and syntax, and in Spanish the difficulty is exacerbated by historical shifts between inflection and parataxis, not all of which have been unidirectional (see below). In this section we shall concentrate on the inflectional morphology of modern Spanish, including periphrastic tenses and forms of address (in the interests of a unified presentation of verb morphology), but postponing the main discussion of clitics and derivational processes to sections 4 and 5 respectively. Analytic Tendencies Because Spanish has shared the general Romance drift away from synthesis towards more analytical forms of expression, and has accordingly swept away whole areas of Latin inflection (notably the entire morphological passive and most of the substantival declension system), the impression is sometimes given that it is a morphologically simple language. This is mis- Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 93 leading. Derivation has always been a favoured and vigorous method of enriching the lexicon, with the effect of creating numerous forms whose overall morphological structure, while reasonably transparent, can hardly be described as simple. An abstract nominal like desaprovechamiento 'negligence' probably consists of six synchronic morphemes, des-aprovech-a-mient-o, with a further historical division -pro(-vech)- fossilised in the root. Most substantives also continue to be inflected for number and gender, and the verb system remains highly inflected in its most frequent paradigms, the trend towards analysis having affected the periphery much more than the core. Numerous complete utterances can be made using a single verb form: llueve 'it is raining', isaldrds? 'will you go out?', sonreiamos 'we were smiling'. Quite complex ideas can be expressed using only a verb and clitic: se escribian 'they used to write to one another'. This concision is largely owed to inflection. Even in the periphrastic forms, where tense, aspect and voice can be 'unpacked' into as many words, the auxiliary remains highly inflected. In the modern language, only interjections and functors (conjunctions, prepositions and a subset of adverbs) are immune to both inflection and derivation. Spanish, consequently, is far from being an isolating language: very few words consist of only one morph. True, in comparison with the formal registers of Latin, the inventory of inflected paradigms is appreciably smaller, but the frequency of the surviving inflections in continuous text has been maintained or even increased. The development of definite and indefinite articles, for instance, means that most Spanish noun phrases have one number and one gender marker more than their Latin counterparts. The 'synthesis index' of Spanish (calculated by dividing the number of morphemes in running text by the number of words) now falls between 1.9 and 2.2 depending on the complexity of the register. Gender and Number in Substantives The canonical structure of Spanish substantives is: (derivational prefix(es)) -I- lexical root + (theme vowel) + (derivational suffix(es)) + (gender marker) + (number marker). Theme vowels only occur in substantives derived from verbs. Gender is overtly marked on a large proportion of nouns, and number on almost all. Only a tiny handful of nouns like crisis 'crisis' consist of a single, unanalysable morph and remain uninfected for plurality; those which do are borrowings. Gender in modern Spanish is a two-term grammatical category with important syntactic functions. It does not correspond to natural gender and must be regarded as inherent and semantically unmotivated for inanimate nouns. (There would, indeed, be a good case for replacing the traditional terms 'masculine', 'feminine' and 'gender' by something less specific, like 'noun classes 1 and 2'.) The sex of animate beings is normally respected — tio : tia 'uncle : aunt', zorro : zorra 'fox : vixen' — but victima 'victim' is Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 94 SPANISH feminine regardless of the sex of the referent, and the names of less common birds, animals and insects are often invariable, so that to refer to a 'male eagle* or 'female beetle' one must say respectively águila macho and escarabajo hembra, not *dguilo or *escarabaja. A small number of nouns are bivalent, like el/la cliente 'customer', but none are neuter. The combination of the so-called 'neuter' article lo with an adjective — as in lo alto 'the highest point', lo peor 'the worst of the matter', lo triste 'the sad part' — is best regarded as a kind of syntactic nominalisation in which lo functions as a [-count] marker. Historically, it represents a Spanish innovation, not a continuation of the Latin neuter gender, which was eliminated before the Old Spanish period from substantives, usually in favour of masculine, though faint traces of it persist in the pronoun system (see p. 95). The mismatch between natural and grammatical gender is compounded in Spanish by a mismatch between grammatical gender and morphological marking. The markers -o and -a show a high degree of correlation with masculine and feminine respectively. Only one native noun in -o is feminine: la mano 'hand', which remains faithful to its etymological gender. (Modern clipped forms like la foto 'photograph', la moto 'motorcycle' and la radio 'radio' are now rallying to its support.) The reliability of -a as a feminine marker is almost as good; eldia 'day' is a latinate exception, but nearly all others are borrowings, such as elguia 'guide' from Germanic, el sofa 'sofa' from Arabic probably via French, and el poeta 'poet' and several others from Greek. A recent source of exceptions, however, has been the popularity of the agentive suffix -ista: nouns like ciclista 'cyclist' and paracaidista 'parachutist', which are obviously bivalent, can be used in Spanish with either el or la, but retain their -a ending when referring to men. The predictability of gender from most other endings is much lower, though it can be greatly improved by reference to etymological and derivational information. For example, of the 200-odd words ending in -ez, over 80 per cent are feminine; but -ez is 100 per cent feminine if it signals an abstract noun derived from an adjective — la delgadez 'slimness', lapalidez 'pallor', la sensatez 'common sense', and 100 per cent masculine in borrowings from Arabic — el ajedrez 'chess', el ajimez 'window arch', el alferez 'sub-lieutenant', el jaez '(ornate) harness'. Whether such information is systematically available to language learners, however, must be doubted. The categories of gender and number are for the most part overtly marked on determiners, demonstratives, pronouns and adjectives of all kinds, as well as nouns. In Castilian, all plural substantives and determiners end in / - s / , though the derivation of plurals from singulars is not quite so straightforward as this implies, since a sizable minority adds the full syllable /-es/ and a few already ending in / - s / remain unchanged. (The /-e-/ of the syllabic plurals is now generally considered to be epenthetic, an analysis sometimes — though less consensually — applied to / - e / as a singular Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 95 ending.) We have already seen the drastic effect on plural marking in those dialects which have lost final / - s / (see p. 85 above). It should be noted, however, that this development does not affect the inflectional character of plural marking: it merely replaces an additive mechanism by an alternation. Determiners Most nouns in connected discourse are preceded by a determiner inflected for number and gender, of which the most common are the articles: Singular: Plural: Definite m. f. el la los las Indefinite m. f. un una unos unas As in most western varieties of Romance (p. 53), the definite articles are semantically attenuated reflexes of the demonstrative ILLE and the indefinites derive from the numerical UNU 'one' (the Spanish masculine is truncated before substantives but appears as uno when used pronominally). Since determiners automatically acquire number and gender from their head noun and usually have unambiguous exponents for these categories, it could be argued that their presence renders the morphological marking of the noun redundant. But there is a minor complicating factor in the apparent mismatch of article and noun which occurs when a feminine singular noun begins with stressed / a - / (sometimes spelled ha-); thus, el agua 'the water' not *la agua, un dguila 'an eagle' not *una dguila, elhacha 'the axe' not *la hacha, and so forth. Historically, these determiners seem to have developed regularly from ILLA and UNA in this specific phonetic environment, and are therefore accidental homonyms of the masculine forms. Whether Spanish speakers now perceive this pattern underlying the surface syncretism, seems very doubtful. If not, perhaps the correct synchronic analysis is to treat la/una as marked feminine singulars, contrasting with el/un, both unmarked for gender. The Spanish demonstratives form a three-term system which correlates with grammatical person: este 'this (of mine)' : ese 'that (of yours)' : aquel 'yonder (of his/hers/theirs)'. One set of forms doubles up for adjectives and pronouns (the latter take an orthographic accent in the masculine and feminine) and the system is essentially identical to its Latin forerunner (p. 53), though with different exponents. As pronouns, the demonstratives each have a further form, esto : eso : aquello, deriving from the Latin neuter. These are used as anaphors for complete sentences or propositions; they can therefore be considered neutral in gender, but no longer specifically 'neuter' since they have no adnominal functions. Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 96 96 SPANISH SPANISH Subject subject Pronouns Pronouns and and Forms Forms of of Address Address In in In addition addition to to the the grammatical grammatical categories categories of of number number and and gender gender found found in the the substantives substantives and and determiners, determiners, the the pronoun pronoun system system encodes encodes person person and and arguably arguably case case (though (though not not as as aa continuation continuation of of Latin Latin case case functions). functions). There There is, is, however, however, both both syncretism syncretism and and asymmetry, asymmetry, as as can can be be seen seen from from the the subsubject presented in in Figure Figure 3.2. ject pronouns pronouns presented 3.2. Here, Here, gender gender is is neutralised neutralised in in the the first and second second persons persons singular. singular. In first and In the the plural, plural, the the /-as/ /-as/ forms forms are are used used to to refer to exclusively exclusively female female sets, sets, and refer to and those those in in l-os/ /-os/ are are therefore therefore better better described 'masculine' (in described as as 'gender-neutral' 'gender-neutral' than than 'masculine' (in much much the the same same way way that that 'masculine' 'masculine' isis better better described described as as the the unmarked unmarked gender gender for for substantives). substantives). Figure 3.2: The Castilian Subject Pronoun System First First person person Singular: Singular: Plural: Plural: yo yo ~ nosotras nosotras nosotros nosotros Second Second person person ~ tu vosotros vosotros Polite Politeaddress address Singular: Singular: us ted usted I • Plural: Plural: ustcdes ustedes vosotras vosotras Third Thirdperson person el 61 I ellos ella I elias ellas The first and third person subject pronouns are common to all varieties of Spanish, but Castilian is sharply distinguished from most others in its treatment of the semantic category of addressee. In the late medieval period, the physical distance encoded in the person category (and in demonstratives) came to be exploited metaphorically as a marker of social distance, the 'polite' address forms ustedlustedes (see p. 119) colligating with third person verb inflections in order to emphasise the differential status accorded by the speaker to the addressee. Quite subtle shades of social relationship can be conveyed by this system. Symmetrical usage by speaker and addressee connotes social equality, plus familiarity if tu is chosen, or respectful formality if us ted is preferred. Asymmetrical usage by consent encodes a relationship of social inequality, even dominance/ subservience. Unnegotiated asymmetrical usage can therefore be employed as an instantly recognisable signal of contempt or aggression. The system remained stable in most of Spain for at least two centuries, but has recently Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 97 been undergoing significant change. The morphological aspects are not affected, but whereas usted was previously the unmarked form, employed in all contexts not specifically judged to be familiar or intimate, it has now become positively marked for deference. Unless there are gross discrepancies of age and status, a stranger will now be addressed in Spain as til. This recent development has the effect of widening the earlier discrepancy in usage between Castilian and other varieties. In grammatical terms, the Castilian system rests on a clash between second person semantics and third person morphology. This clash is resolved in west Andalusian and Canary Island dialects by colligating ustedes with second person morphology, as in ustedes sois 'you are'. In Latin America, the position is more complicated. Almost all regions lack the European 'familiar' plural vosotros (whether they have eliminated the form, or never had it, is disputed). Instead, ustedes is used, with third person inflections, as a generalised plural. In the singular, there is more diversity. Standard Mexican Spanish uses tu opposing tu hablas 'you speak' in the singular to ustedes hablan in the plural. Standard Argentinian uses as its generalised singular vos (which in medieval Spain had been used as a polite singular, just as modern French vous), colligated with inflections which derive from the second person plural but are no longer identical to their European counterparts, hence vos pones 'you put', vos andds 'you walk'. Between Mexican and Argentinian, a wide array of permutations is possible, colligating pronouns with inflections which are historically both singular and plural, sometimes even blends. The stable voseo of Argentina is not a recent phenomenon; its roots must be sought in the colonial period, and archival research has revealed that it was well established in educated Buenos Aires usage by the beginning of the nineteenth century. This suggests that voseo, which is found in Uruguay, Paraguay and most of Central America as well as Argentina, and preserves the essentials of fifteenthcentury peninsular usage, may originally have been the principal, if not uniform, model throughout Latin America. The pattern now found in Mexico and most of the Caribbean probably represents the historical overlay of later Castilian changes, diffused through better contacts with the Viceroyalty, and meeting more or less resistance as they filtered southwards. Verbal Inflection We turn now to a discussion of the simple tense-forms of the verb. In the standard language verbs are traditionally said to belong to one of three conjugations, with infinitives in -ar, -er and -ir. The -ar group, deriving from the Latin first conjugation in -ARE, is by far the largest and the one which accommodates almost all new coinings (compare alunizar 'to land on the moon' with French alunir, p. 223). The distinction between the -er Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 98 SPANISH and -ir patterns is more apparent than real: aside from the future and conditional paradigms (which necessarily diverge since they take the infinitive as their stem) their endings are identical in all but four instances. We shall therefore distinguish only two basic conjugations for regular verbs, as set out in the chart of simple tense-forms. In Castilian, as in Latin, each paradigm consists of six forms representing three grammatical persons in both singular and plural (most American varieties, as we have seen, have only five forms). In general, all The Simple Tense-forms of Regular Verbs, Showing the Stress and a Possible Morphological Analysis Conjugation I: tomar 'to take' (a) Present Indicative tom-Ø-o tóm-a-s tóm-a-Ø tóm-á-mos tóm-á-is tóm-á-n Indicative com-Ø-o cóm-e-s cóm-e-Ø cóm-é-mos cóm-é-is cóm-é-n Subjunctive tom-e-Ø tóm-e-s tóm-e-Ø tóm-é-mos tóm-é-is tóm-é-n (b)Imperfect Indicative Subjunctive (1) or tom-á-ba-Ø -á-se-Ø -a-se-s tom-á-ba-s tom-á-ba-Ø -á-se-Ø tom-á-ba-mos -á-se-mos -á-se-is tom-á-ba-is -á-se-n tom-á-ba-n Conjugation II: comer ‘to eat' (2) -á-ra -Ø -á-ra -s -á-ra -Ø -á-ra -mos -á-ra -is -á-ra -n Subjunctive com-a-Ø cóm-a-s cóm-a-Ø cóm-á-mos cóm-á-is cóm-a-n Indicative Subjunctive (1) or com-i-a-Ø -ié-se-Ø -ié-se-s com-i-a-s com-i-a-Ø -ié-se-Ø com-i-a-mos -ié-se-mos com-i-a-is -ié-se-is -ié-se-n com-i-a-n (c) Preterit or simple past (indicative only) tom-Ø-6 (?e = a>i) tom á-ste tom-Ø-6 (?6 = á+u) tom-6-mos tom-á ste-is tom-á-ro-n com-Ø-i (?f - i+i) com-i-ste com-i-6 com-i-mos com-i-ste-is com-ié-ro-n (d)Future indicative (all verbs) tom-a-r-é tom-a-r-ás tom-a-r-á tom-a-r-emos tom-a-r-éis tom-a-r-án Conditional I(all verbs) com-e-r-ia com-e-r-ias com-e-r-ia com-e-r-iamos com-e-r-iais com-e-r-ian (2) -ié-ra-Ø -ie-ra-s -ie-ra-Ø -ie'-ra-mos -ie-ra-is -ie-ra-n Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 99 six forms are distinct, though there is some syncretism in first and third persons singular, and more in dialects which have lost final / - s / . As already noted (see p. 87), stress operates functionally to differentiate otherwise identical forms. The unmarked paradigm is the present indicative and the unmarked person the third singular, which is the morphological shape assumed by the handful of verbs that do not accept animate subjects (nieva 'it is snowing', acontecio 'it happened'). It is useful to distinguish a 'theme vowel' after the lexical stem, /-a-/ for the first conjugation and for the second /-e-/ or /-i-/, in a rather complicated phonological distribution. It can then be seen that the distinction between the present indicative and subjunctive rests on a reversal of the theme vowel. The order of morphemes is fixed: (derivational prefix(es)) + lexical stem -f theme vowel + tense marker (sometimes including an empty morph) -f person marker. Some forms, however, have fused in the course of history and a neat segmentation is not always possible. The preterit is the most difficult paradigm to analyse, since the theme vowel is sometimes indistinguishable, and segmenting the second and third person plural markers in the regular way, /-is, -n/, leaves an awkward residue which occurs nowhere else in the system. The future and conditional pose a rather different problem: both have evolved during the history of Spanish (see below) from combinations of the infinitive with either the present or imperfect of the auxiliary haber 'to have', and despite considerable phonetic reduction the 'endings' still contain traces of this verb's lexical stem. This secondary derivation explains the identity of the conditional endings with those of the second conjugation imperfect. Spanish is in the unusual position of having alternative forms for the past subjunctive, neither of which is a reflex of the original paradigm. The -se series derives from the Latin pluperfect subjunctive, and the -ra from the pluperfect indicative (see p. 46). In northwestern dialects and parts of Latin America, -ra is still occasionally found as a pluperfect. In standard Spanish, the two forms are not quite interchangeable: in the 'attenuating' sense quisiera i should like' and debiera 'I really ought' cannot be replaced by the -se counterparts, and elsewhere their distribution may be determined by considerations of symmetry or by sociolinguistic factors. Limits of Regularity By the strictest criteria, almost 900 Spanish verbs are irregular in one or more of the simple tense-forms (excluding those which undergo merely orthographic alternations, discussed on p. 90). This disconcerting figure includes a very few with anomalies in their endings, for instance the empty morph -y which appears in doy ‘I give' and only three other verbs (see the chart of irregular verbs). All the others are subject to stem alternations, with varying degrees of predictability. Over half the total are 'radical Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 100 SPANISH ser 'to be' (a) Present indicative soy eres es somos sois son estar 'to be' haber 'to have ' tener 'to have ' ir 'to go' estoy estas esta estamos estais estan he has ha hem os habeis han (b)Present subjunctive (endings regular, same stem este sea haya estes seas hayas este sea haya seamos estemos hayamos esteis seais haydis esten sean hayan tengo tienes tiene tenemos tenuis tienen throughout) tenga tengas tenga tengamos teng£is tengan voy vas va vamos vais van vaya vayas vaya vayamos vayals vayan (c) Imperfect indicative (endings regular, same stemi throughout) tenia habìa estaba era tenias habìas estabas eras habìa estaba era tenìa eramos tenìamos habìamos estábamos erais tenìais habìais estabais tenìan habìan estaban eran iba ibas iba ìbamos ibais iban (d)Future indicative (endings regular, infinitival stem throughout) tendre habre sere estare ire (e) Conditional (endings regular, infinitival stem throughout) estaria tendrìa habrìa serìa irìa (f) Preterit indicative fui fuiste fue fuimos fuisteis fueron (endings slightly irregular, preterit stem throughout) estuve hube tuve fui estuviste hubiste tuviste fuiste estuvo hubo tuvo fue estuvimos hubimos tuvimos fuimos estuvisteis hubisteis tuvisteis fuisteis estuvieron hubieron tuvieron fueron (g)Imperfect subjunctive (endings regular, preterit stem throughou t) (l)fuese estuviese hubiese tuviese fuese (2)fuera estuviera tuviera hubiera fuera changing', of the three types discussed earlier (p. 88); some others, like huir 'to flee', insert a glide under predictable conditions. A significant minority retain a Latin stem alternation which is no longer predictable or productive, contrasting the preterit and the non-present subjunctive paradigms with all others; these verbs, as can be seen from the chart, have Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 101 slightly different endings in the preterit, with an alternating stress pattern. A number of verbs, some of which are otherwise regular, preserve a Latin anomaly in the past participle — abrir 'to open' : abierto < APERTUM, romper 'to break' : roto < RUPTUM; while the common irregular participles like hecho 'done' < FACTUM show no sign of change, others seem likely candidates for analogical replacement, as has happened with torcer 'to twist', now conjugated with torcido though its old participle tuerto survives as an adjective meaning 'one-eyed/squint-eyed'. Some twenty verbs of conjugation II modify the infinitival stem used as the basis of the future and conditional, mostly as a regular phonological adjustment following the loss of the theme vowel before a stressed inflection, as in saldre < salir + é 'I shall go out', valdria < valer + ia 'it would be worth'. Finally, a handful of very frequent verbs are totally eccentric and even undergo stem suppletion. One class, amounting to some 200 including compounds, deserves special mention. Polysyllabic verbs which end in -cer or -cir preceded by a vowel, like conocer 'to know' or relucir 'to flaunt', have an extra velar consonant before non-front vowels, conozco being pronounced [ko-'naO-ko] in Castilian and [ko-'nos-ko] in seseante districts of Andalusia and throughout Latin America. The intriguing question is: where does the velar come from? Is it part of the underlying stem but lost before front vowels? Or is it epenthetic, and if so under what conditions? The first answer is historically correct: ail these verbs contain an originally inchoative infix -sc(for French, see p. 223), whose velar regularly palatalised before a front vowel and assimilated to the preceding sibilant. But it seems unlikely that contemporary speakers recapitulate this process to produce the less frequent of the two alternants. If the velar is regarded as epenthetic (though phonetically unmotivated), it remains predictable in Castilian but only by reference to the phoneme / θ / . In seseante dialects, which lack the /θ : s/ opposition, the alternation is unpredictable: speakers cannot know from the phonological structure that reconocer 'to recognise' [re-ko-no-'ser] requires [-k-] while recoser 'to sew up' [re-ko-'ser] does not. They must, in other words, learn the alternation as an inherent lexical feature of the verb. Periphrastic Forms and Auxiliaries In addition to its simple paradigms, Spanish is particularly well endowed with periphrastic forms, more so than any other standard Romance lan guage. Usually, these consist of an inflected auxiliary followed by a nonfinite form of the lexical verb (an infinitive or participle), but more complex combinations are also possible. Virtually all are Romance creations, though some embryonic models are attested in Latin (pp. 56-8). The most farreaching innovation was the compounding of HABERE, originally meaning 'to possess', with a past participle, HABEO CENAM PARATAM first meant 'I have the supper here, already prepared', but with increased use and a Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 102 SPANISH change of word order, it soon came to mean simply ‘I’ve prepared the supper’The new construction provided a powerful model: in principle, any paradigm of HABERE could be combined with the past participle to make a new tense-form. This remains true in modern Spanish: all eight simple paradigms of haber, including the almost extinct future subjunctive, can be used as auxiliaries. Although the periphrases with haber were flourishing in Old Spanish, they could only be used with transitive verbs, a direct consequence of their etymology; intransitives were conjugated with ser (see p. 57). It was only at the end of the fifteenth century that haber ousted ser for all verbs, and the past participle became invariable. Alongside the now fully grammaticalised periphrases with haber, however, Spanish has had since its earliest texts an alternative construction making use of tener < TENERE 'to hold' closely reminiscent of the prototype: tengo preparada la cena, with agreement, means the same as the Latin expression from which we set out. (For the use of cognate term Portuguese, see p. 150.) The chart of irregular verbs, detailing the most common auxiliaries, shows not only two verbs 'to have' but also two verbs 'to be', a notorious difficulty for foreign learners of Spanish. At some risk of oversimplification we shall say that ser is the normal copula, denoting inherent qualities, while estar focuses on resultant states; compare la pimienta es picante 'pepper is hot' (inherently) with la sopa está friá 'the soup's cold' (because it's cooled down). Both verbs can be used as auxiliaries, in conjunction with a past participle, to make analytic passives. This results in a plethora of forms, since any paradigm of ser or estar can be used, including those which are already composite. Nor are the two passives synonymous: ser denotes the action or process, as in el dinero ha sido robado (por un atracador) 'the money has been stolen (by a gangster)', whereas estar denotes the subsequent state, as in la tienda estaba abierta 'the shop was open' (because it had been opened). Estar also combines with a present participle to create a range of progressive forms. In turn, these may combine with other periphrases, without grammatical restriction, but nevertheless with increasing markedness, so that three-term verbs like habia estado andando 'I'd been walking' are not frequent, and monsters like ha estado siendo construido 'it's been being built' are usually avoided in compassion for the listener. Evolutionary Trends To conclude this section on morphology, we should pose the twin questions: how secure is inflection? and could the elaborate verb inflection system be reduced or eliminated? So drastic a change is inconceivable in the near future, and it would be wrong to assume that Spanish is necessarily evolving along the same pathways as other Romance varieties, notably French, where inflection has been reduced. Indeed, in the course of its history, Spanish has created new morphology. One example is the adverbial ending -mente, deriving from the ablative of the feminine noun Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 103 MENS m ' i n d / m a n n e r ' used with a feminine adjective, thus STRICTA MENTE > estrechamente 'narrowly'. In modern Spanish -mente is no longer perceived as a noun (its congener miente 'mind' is now confined to a few idiomatic expressions), but it can still be detached from its host adjective in order to avoid awkward repetition: lenta(*mente) y cuidadosamente 'slowly and carefully'. Even clearer examples are the future and conditional paradigms, both general Romance innovations (see p. 57), which in Old Spanish were still separable into their component parts — witness conbidar le ien de grado 'they would willingly invite him', Mio Cid 21 — and in Portuguese have still not fully coalesced (see p. 150). Although the new future paradigm is, in its turn, tending to be replaced in colloquial Spanish by a periphrasis (voy a ver 'I'm going to see/ I'll see') which better accords with general analytic tendencies in the language, the conditional paradigm is under no threat of substitution — on the contrary, it appears to be extending its range of functions. Against this, we must set a theoretical prediction and a tentative observation. In typological theory, VO syntax and suffixal inflection may be viewed as cross-category disharmonies. On a syntactic level, there is no doubt that Spanish has been moving towards VO structure and has attained a reasonable consistency (see section 4); its verbal inflections constitute a prominent exception. Their demise, if it did come about, would have to be the consequence of a major syntactic reorganisation. Such a reorganisation occurred in the sparsely documented period between Latin and Old Spanish, when the development of determiners and the increased use of prepositions made most of the case system redundant. Something similar happened in Middle French (p. 231), when the use of subject pronouns became obligatory and the verb endings, now redundant, began to coalesce and disappear. Spanish, as we have seen, shows little syncretism in its inflections and, unlike French, rarely needs subject pronouns to avoid syntactic ambiguity, though they are regularly used for emphasis and contrast. But any move to increase the use of subject pronouns — and there is some evidence that this is happening in colloquial registers — would undermine the need to preserve number and person inflections in the verb, and could thereby threaten the entire inflectional edifice. 4 Syntax Typological Consistency On most of the criteria favoured by typological theory, modern Spanish is a consistent VO language. Briefly: in simplex sentences objects or complements follow the main verb; noun phrase relationships are expressed exclusively by prepositions; genitive constructions take the form of prepositional phrases and always follow their head noun; the standard follows Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 104 SPANISH the comparative; most adjectives, and all attributive phrases and relative clauses, follow their head noun; most adverbs follow the verb they modify; auxiliaries are frequent and always precede the lexical stem even when they are themselves composite; quantifiers and negatives precede the item they qualify and have only forward scope; interrogative words are always phrase-initial. Needless to add, there are some complications. We have already noted one major discrepancy with VO typology: the vigour of suffixal inflection in the verb system, a vigour which has been fed rather than starved by the evolution of auxiliaries. We shall examine below some apparent inconsistencies both in the effects of adverbial scope and in the operator : operand relations of derived lexical forms and compounds. Despite these reservations, there can be no doubt that, during the period of its documented history, Spanish has evolved in ways which increase, rather than disturb, its typological consistency. While the chief distributional facts of Spanish syntax are relatively uncontentious (when due allowance is made for some regional variation), it is nevertheless difficult to present a theoretically neutral description of some core features which are susceptible of competing analyses. The key difficulties are: word order, constituency and concord. We shall assume a model which, like typological theory, recognises syntactic relations as primes. It is clear, for instance, that grammatical mobility is an attribute of subjects in Spanish, not a general property of noun phrases. As we shall see, subject and object phrases are formally differentiated in ways which obviously relate to their sentential functions. The Nominal Group Subjects The minimal overt structure of a subject noun phrase in Spanish is a proper noun or single personal pronoun (see Figure 3.2 above). For subject NPs with common nouns as their heads, the structure is: determiner + (numeral) + (adjective or conjoined adjectives) + noun + (adjectival modifier(s)). In Spanish, the determiners — articles, demonstratives and (unlike Italian, pp. 54,298) possessives — form a mutually exclusive set, only one being possible in prenominal position. Articles always precede the noun and have done so quite consistently since the earliest documented stages of the language. Demonstratives and possessives may, however, be used as postposed adjectives, in which case the preposed article is retained, so ese libro 'that book (of yours/ near you)' alternates with el libro ese which is more colloquial and sometimes pejorative. Most of the possessives have several alternants, with a short form, inflected only for number and restricted to prenominal position, opposed to a longer form, inflected for both number and gender, and used as a postnominal adjective. Hence: mi hermano 'my brother', mi hermana 'my sister', misparientes 'my relatives', Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 105 un libro mio 'a book of mine/one of my books', unas sobrinas tuyas 'some of your nieces'. When used pronominally, the demonstratives stand alone (embellished by an orthographic accent) — este 'this man', aquella 'that woman'; but the possessives must be combined with an article — el mio 'mine (of masculine singular possessions)', las nuestras 'ours (of feminine plural possessions)'. Quantifiers and negatives form a special class sharing some characteristics with adjectives and others with determiners, and therefore cut across the schema given above. Both precede the noun they modify (a possible exception is discussed below), but a negative is not compatible with a determiner — ningunos libros 'no books' (*los/*unos ningunos libros), whereas a quantifier can cooccur with a definite determiner — (tus) muchos amigos, '(your) many friends', (los) pocos drboles 'the (few) trees'. Combinations of quantifiers and indefinite articles are not usually permissible — *unos muchos '*some many', *un cada '*an each'; but unos pocos 'some/a few' is not only perfectly grammatical but also very frequent. Faced with this apparent exception, some linguists have reanalysed unos/unas as a special kind of quantifier, rather than as the plural of the indefinite article unluna. Adjectival Modifiers and Relatives In Spanish, adjectives but not adjectival phrases or relative clauses may precede the head noun. The handful of quantifiers and negatives already mentioned are the only adjectives to be invariably fixed in prenominal position. Some common adjectives have clearly differentiated meanings depending on position: un buen hombre 'a good chap' contrasts with un hombre bueno 'a morally upright man', as does un viejo amigo 'a long-time friend' with un amigo viejo 'an elderly friend'. The semantic and pragmatic conditions for adjective placement have long been a matter of debate in Spanish linguistics and a limited consensus has emerged that prenominal adjectives indicate known, expected or figurative qualities whereas those following the noun convey new, unexpected or distinguishing information; compare las verdes hojas 'the green leaves' with las hojas azules 'the blue leaves'. But semantic generalisations about the content of individual adjectives are quite often falsified by usage in extended contexts. A recent proposal based on discourse analysis claims that the relative order of adjective and noun is itself meaningful: postnominal order denotes contrast, the establishment of a difference, whereas prenominal order merely provides a characterisation without implying any contrast. It should be noted that this analysis works well not only for conventional adjectives, but also for any pairs of substantives, like obra maestra 'masterpiece', in which the first slot is reserved for the element being characterised and the second for the characteriser. Relative clauses in Spanish must follow their head noun or dummy head Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 106 SPANISH and be introduced by an overt relativiser. Que, the unmarked complementiser, is also the particle most often used to introduce relatives. During the documented history of Spanish, que has been increasing its domain at the expense of quien(es) 'who', and the supposed prehistorical opposition of inanimate que to animate quien has now been so far undermined that combinations like elhombre quien... 'the man who ...' sound over-formal and stilted in many contexts. For many Spanish speakers, however, it remains ungrammatical to use que in oblique relatives with human antecedents; hence el chico a quien ... (*a que ...) 'the boy to whom ...', el abogado con quien ... (*con que...) 'the lawyer with whom ...'. To indicate a possessive relative, de quien 'of whom' may be used, but much more common is cuyo which derives from the invariable Latin genitive pronoun CUIUS, but has been reanalysed as an adjective and now agrees with the item possessed, hence el chico cuyas hermanas ... 'the boy whose sisters ...'. This historical reanalysis allows modern Spanish to 'access' prepositional objects of possessives for relativisation much more easily than can be achieved in some other Romance languages: el chico con cuya hermana hablabas 'the boy whose sister you were talking to'. Objects Many of the characteristics of subject noun phrases, including the behaviour of their adjectival modifiers, carry over to object or complement phrases, but there are some important exceptions. In particular, for nonsubject NPs, determiners are distributed according to semantic criteria: there are few purely syntactic constraints on their presence. For mass nouns no determiner is used: compre vino/arroz 'I bought (some) wine/rice'; there is no equivalent of the 'partitive' article obligatory in French in comparable environments (see p. 227). Singular count nouns do require a determiner — compre un libro/el libro que querias 'I bought a book/the book you wanted'; but in the plural it may be omitted, producing a slight semantic difference between compre libros (= some, an indetermiate number) and compre unos libros (= a few, a small number of specific books). In the case of possessive pronouns, the article is obligatory for direct objects — jdale el tuyo! 'give him yours', but is usually omitted in complements — este lapiz es mio 'this pencil is mine' — resulting in a neutralisation of the potential distinction between adjectival and pronominal complements. A further peculiarity of Spanish object NPs is illustrated by vi a Carmen 'I saw Carmen' and vi a tu primo 'I saw your cousin', as opposed to vi la calle 'I saw the street' or vi (unas) montanas 'I saw (some) mountains'. When the object refers to a particular human being, it is obligatorily introduced by the preposition a, popularly known as 'personal a\ This construction is by no means confined to humans; it is common when the referent is an animal, a place name or country, or even an inanimate object if there is any possibility of confusion with the grammatical subject, hence: Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 107 reconocia a Roma 'she could recognise Rome', la rata cazó al gato 'the rat hunted the cat', la bicicleta doblo al camion 'the bicycle overtook the lorry'. This usage has been increasing during the documented history of Castilian, where a now seems to function as a generalised object marker, with the important proviso that the object so marked must be particularised or referential. Preserving the distinctiveness of subjects and objects is an evident desideratum in Spanish syntax (and, as we shall see, one of the main guarantors of relatively free word order), but its maintenance is not achieved without wider repercussions. In particular, the 'personal a' construction has the effect of converting many object NPs into prepositional phrases and, since a is also the preposition used to introduce datives, of removing any overt distinction between direct and indirect objects for the great majority of animate NPs; witness vi a Juana 'I saw Jane' and di a Juana el recado 'I gave Jane the message'. Object Pronouns and Clitics The grammar of object pronouns in Spanish is notoriously complex and also liable to considerable regional variation; we shall be able to present only a schematic outline here. The pronouns themselves are of two morphological types: disjunctive free forms and clitic particles bound to the verb (though the clitics also derive historically from free pronouns). Object pronominalisation in Spanish never consists merely of substituting a free pronoun for a full NP in the identical syntactic frame. In many sentences two coreferential pronouns occur, one disjunctive and the other clitic; of these, it is the clitic which is obligatory. The inclusion of the matching disjunctive pronoun signals emphasis, contrast or diasmbiguation. So, the neutral pronominal equivalent of vi a Juana 'I saw Jane', is la vi I saw her', with an emphatic or contrastive variant in la vi a ella T saw HER'; but * vi a ella is ungrammatical, unlike the equivalent structure in Italian (p. 290). So-called 'clitic doubling' is widespread in Spanish and is not confined to examples of the above type. Full NPs with human referents also commonly occur with a clitic: la vi a Juana 'I saw Jane', le dije a Miguel que ... 'I said to Michael that ...'. This results in a skewed manifestation of 'doubling'. When the direct or indirect object is a proper noun or definite phrase with a human referent (and sometimes, by extension, any animate referent), doubling consists of inserting a clitic 'copy' of the object, a process which is optional but often preferred. But when the object is human and pronominal, the clitic is obligatory, so that doubling consists of optionally including the disjunctive pronoun. Doubling, in both contexts, is clearly gaining ground, and for a few verbs has become almost mandatory: for many speakers, gustar 'to please', now requires a clitic even when its indirect object is a human proper noun — le gustan a Maria las cerezas (*gustan ...) 'Mary likes cherries' (literally 'to-her are-pleasing to Mary the cherries'). On the other hand, doubling is not permissible when the Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 108 SPANISH object NP is inanimate. The third person deictic clitics lo, la, los, las cannot cooccur within a simplex sentence with the full NP, hence compro el libro 'she bought the book', lo compro 'she bought it', but not *lo compro el libro. (In this respect, Spanish is rather unlike French; see p. 236.) We noted above that the use of 'personal a' precludes an overt distinction between direct and indirect objects with animate referents. What is true of full NPs is also true of disjunctive pronouns and furthermore of most clitics: compare me vio a mi 'she saw ME' with a mi me dijo que ... 'she said to ME t h a t . . . ' . Only in the third person clitics is an overt distinction made between direct and indirect objects (or 'accusatives' and 'datives' in different terminology): la vi (a ella) T saw her' versus le dije (a ella) que ... i said to her that ...'. This distinction, though potentially useful, is unique within the pronoun system and is now being subjected to analogical levelling. Since, however, there are competing targets for the process, levelling may well result in regionally differentiated patterns. The reason is that the third person system inherited from Latin distinguishes direct from indirect, but not always animate from inanimate — a dichotomy which is very important in other parts of the grammar. La vi, without further context, is ambiguous between 'I saw her' and T saw it' (any feminine-gender inanimate object). The changes in progress are schematised in Figure 3.3. A Castilian innovation using le < dative ILLI as the animate masculine direct object clitic, has spread rapidly and in Spain is now virtually standard. This usage, called leismo, has not penetrated to some remoter parts of Spain or to most of Latin America, which remains faithful to the etymological model, called appropriately loismo. In Castilian, the logical extension of lo > le is for plural les to oust los as the animate masculine direct object; this too is happening, though a little more slowly. A further development, called laismo, now clearly established in parts of Old Castile but still stigmatised elsewhere, uses la and las to mark animate feminine indirect objects as well as direct, hence la di el libro 'I gave her the book'. If both leismo and laismo eventually become standard in the prestige dialect, the effect will be to eliminate all overt trace of the Latin accusative : dative opposition. Clear differentiation of gender will have been a target attained at the expense of case. But the resulting system will still not be symmetrical, since the animate : inanimate distinction will be overtly marked for masculines, but not for feminines. Therein may lie the seeds of a further readjustment in the future. Cliticisation of pronouns to the verb is by no means a recent process in Spanish; it appears well established at the time of the earliest texts. What has changed is the position of clitics with respect to the verb. Enclisis was frequent in older stages of the language, and even in the late nineteenth century it was regarded as elegant (albeit slightly precious) in novelistic style to preserve enclitic reflexives especially in the preterit; so, for Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 109 Figure 3.3: Variation in the Third Person Clitic System non-reflexive clitic object direct masculine referent inanimate Late Latin indirect feminine referent masculine referent feminine referent animate ILLO ILLO ILLA ILLI ILL! Conservative Spanish (loismo) lo lo la le le leismo lo Ie la laismo lo lo la le la leismo + laismo lo le la le la Ie le Note: An overt distinction between animate and inanimate operates only where shown. For plural forms, add / - s / throughout. All the above items are mutually exclusive; no two /-forms may co-occur in any variety of Spanish (see p. 110). instance, Galdos frequently writes, in sentence-initial position, sentose 'he sat down' instead of the more prosaic se sentó, and levantóse 'he stood up' instead of se levanto. Reference grammars of contemporary Spanish customarily list three conditions under which enclitic position is maintained: in the positive imperatives — jddmelo! 'give it to me', and after infinitives and present participles — voy a verlo 'I'm going to see it', sigue cantdndolo 'he goes on singing it'. In fact, in colloquial Spanish, the last two examples would be lo voy a ver and lo sigue cantando respectively. Increasingly too, clitics 'climb' from a lower clause to the front of the main verb, sometimes hopping over an intermediate element; compare formal quiero hacerlo 'I want to do it' and tiene que traermelo 'he must bring it for me' with colloquial lo quiero hacer and me lo tiene que traer. Despite the frequency of such examples, there are a number of contexts in which climbing is not possible, some apparently determined by lexical properties of the individual verb, others relatable to more general features like the presence of an intervening preposition. For instance, creer 'to believe' and sentir 'to Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 110 SPANISH regret' do not normally accept climbing from a lower clause: creo haberlo dicho/*lo creo haber dicho 'I believe I said it', siento haberlo hecho/*lo siento haber hecho 'I'm sorry I did it'; while the linking prepositions en and con seem regularly to block climbing: insisto en hacerlo/*lo insisto en hacer 'I insist on doing it', sueño con verlo/*lo sueño con ver 'I dream of seeing it'. While the overall frequency of clitic climbing seems to be increasing in Spanish, there is some tentative evidence that the number of individual sites permitting the movement has declined since the medieval period, as has certainly been true of French; again, there seem to have been competing targets for analogical levelling. Curiously, voice is the verbal category with which pronominals have been most closely linked during the history of Spanish. The connection is essentially due to the cliticisation of the reflexive pronouns, a development which seems likely to result in the emergence of a new set of medio-passive paradigms. In modern Spanish reflexivity can be marked unambiguously on disjunctive pronouns by the addition of mismo 'self, same', as in (a) ti mismo 'yourself; but most of the clitics, which cannot co-occur alone with mismo, are ambiguous in this respect (as also between direct and indirect interpretation). Reflexivity is distinguished, if at all, only in the third person se, which contrasts with the non-reflexive deictics in /- (see Figure 3.3), but itself neutralises the direct : indirect opposition, and also number and gender. Recently, se and its congeners in other Romance languages have been the focus of intense linguistic debate. The problem is whether se should be treated as one single morpheme or a set of homophonous forms. Traditional accounts distinguish four or more functions of se which can be traced, more or less directly, to Latin SE, SIBI, or a fusion of the two: a true reflexive pronoun — se lavo 'he washed himself; a reciprocal marker — solian escribirse 'they used to write to one another'; a passive marker — el congreso se inauguró 'the congress was opened'; and an impersonal marker — se habla inglés 'English spoken'. In addition, Spanish uses a se, known to have a different historical origin, as a means of circumventing the prohibition, dating from the earliest texts, on the cooccurrence of any two /clitics; hence se lo dio 'she gave it to him', not *le lo dio. (This strategy for avoidance is peculiar to Spanish, though the prohibition itself is not.) The difficulty for synchronic analysis is that several of the above functions appear to be semantically compatible, but Spanish never in practice permits more than one se per verb. The combination, for example, of an 'impersonal' se with a 'lexically reflexive' verb is ungrammatical — *se se esfuerza por ... 'one struggles to ... ', as are many other apparently reasonable pairs. If se were only one morpheme, the problem would not arise; but can such disparate meanings be reconciled? Two accounts are now available which solve most of the problems. In one, se is viewed as a pronoun with very little inherent meaning ('third person, low deixis'), which acquires stg- Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 111 nificance from contextual inferences. In the other, se is seen as part of a new medio-passive paradigm, its third person impersonal use paralleling that of Latin: VIVITUR = se vive = 'one lives'. In neither treatment is se a reflexive pronoun! The Verb Tense and Aspect Spanish verbs, as we pointed out above (p. 93), are highly inflected and can often stand alone to form a complete utterance. As such, they are necessarily the principal carriers of tense, aspect and modality, though this information can be devolved upon an auxiliary and supplemented from various adverbial structures. If the 'core' system of Latin is viewed as the interaction of three time values (past, present, future) with two aspects (imperfective, perfective; see p. 56), then a substantial continuity of functions is discernible in modern Spanish, despite some notable modifications in morphology. The morphological changes have, however, imposed considerable strain on the semantic relations simply by making available more candidates for core membership than can be accommodated within the Latin six-cell matrix. Especially problematic in this respect are the past and conditional paradigms. The emergence of HABERE as an auxiliary not only provided a neat morphological marking of perfective aspect, which in Latin had been dependent on a stem alternation, but also resolved the bivalency of the Latin 'perfect' as between present perfective and simple past (aorist) functions (see pp. 56 and 228). Thus vivi in the sense of T have seen' was replaced by HABEO *VISTU > he visto, leaving vi to denote unambiguously 'I saw/I caught sight of. It would be difficult, in modern Spanish, to argue for the exclusion of either of these forms from the core system, though the inclusion of both requires the recognition of a third aspect (punctual) which is nevertheless only opposed to the other two on the past time axis. (The so-called 'past anterior tense', as in hube visto < HABUi *VISTU, which is restricted to a subset of temporal adverbial clauses in the literary language, cannot be considered a member of the core system; nor does it systematically mark punctuality.) The resulting asymmetry has led some commentators to suppose that Spanish might follow French (p. 228) or northern Italian (p. 300) in extending the domain of the periphrastic paradigm to the point where it ousts the simple past from everyday speech and recreates the bivalency of Latin; but the Spanish evidence is difficult to interpret. The essential denotation of forms like he visto is 'completion of an action or process which remains relevant at the moment of speaking', but there are well known regional differences in usage according to the time lapse between completion and moment of speaking. In Castilian and central peninsular dialects the tolerated time lapse is longer, so that he visto Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 112 SPANISH is semantically compatible with time adverbials like esta manana 'this morning' and, for some speakers, even with ayer 'yesterday'. In other peninsular dialects such collocations are not found, and in many Latin American varieties, notably Mexican, the periphrastic forms are reserved for immediate past time, so that ayer he visto... would simply be ungrammatical. The obvious conclusion that Castilian is the innovating variety is, however, checked by very recent observation of the educated norm, in which the preterit vi is increasingly being used to refer to the immediate past. It is too early to tell whether this represents a temporary aberration, or a genuine trend imitated from a newly prestigious conservative cognate. Another source of tension in the core system has been the ambiguous status of the conditional. A plausible pathway for its incorporation is sketched below in relation to French (p. 229). If the absolute time values 'past: present: future' are reconceptualised as relative values 'anteriority : simultaneity : posteriority', this opens the way for a reanalysis of aspect in terms of 'now : then' axes. The new framework allows the conditional veria 'I should see' to be incorporated as the signal of 'posteriority' on the 'then' axis, in the same way that the future vere 'I shall see' represents 'posteriority' on the 'now' axis. This accurately reflects the distributional match of the two forms in indirect and direct speech respectively. If we take the argument one stage further and recognise a dimension of 'anticipated anteriority', we can also accommodate habria visto 'I should have seen' on the 'then' axis, to match habre visto 'I shall have seen' on the 'now'. Again, the distribution of the two forms in indirect and direct speech runs parallel. The difficulty with this neat analysis is that conditionals carry not only temporal-aspectual meaning, but also modality (as, to a lesser extent, do futures; see below). One of their major functions is to evoke a hypothetical or unreal state of affairs. This usage appears to be a more recent historical development of the 'basic' temporal-aspectual meaning, and is probably due to a general tendency for markers of 'posteriority' to acquire modal connotations simply because the state to which they refer is unverifiable at the moment of speaking. In this function, Spanish conditionals are progressively encroaching on the domain of the subjunctive and have therefore a claim to be considered as an independent mood, but one which in relation to the indicative has a very restricted range of tense-forms. Some developments in the 'peripheral' system are also noteworthy. In the core paradigms, imperfective aspect neutralises the potential distinction between habitual and durative meaning, so that hacia can mean either 'I was doing' or i used to do', depending on general context or the presence of a disambiguating adverbial like en aquel tnomento 'at that moment' or todos los dias 'every day'. So much was true of Latin, but Spanish has now evolved symmetrical periphrases to differentiate the meanings, respectively: estaba haciendo and solia hacer (literally 'I was accustomed to do'). The composite forms are not only more specific in aspectual meaning than Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 113 their simple counterparts; they are also more clearly differentiated in time reference, so that, for example, the present durative cannot combine with a future adverbial to refer to immediate futurity — mahana voy a Londres Tm going to London tomorrow', but not *manana estoy yendo. In connection with the development of passives, we alluded earlier (p. 102) to the slow emergence of a stative aspect using estar which now seems to have attained a stable distinctiveness in relation both to the ser-passive and to the reflexive medio-passive. These latter are now principally differentiated by register: ser-passives, though common in journalistic and technical writing, have been virtually ousted from speech to the advantage of the clitic forms. Nevertheless, there is a syntactic difference in that, for most speakers, only the ser-passive can cooccur with an explicit agentive phrase; and also a semantic difference in that the deontic modality of the Latin passive now attaches only to the reflexive forms: eso no se hace 'that isn't done'. There may be a connection between the two distributions, deontic modality being inferred from the absence of agency: 'that isn't done (by anyone)' implying 'that mustn't be done'. Modality and Mood As we have seen, modality in Spanish is conveyed by various mechanisms, but the language is not usually considered to possess a discrete set of modal verbs. The chief exponents of possibility and volition are respectively poder 'to be able' and querer 'to want', which have only minor morphological irregularities and share syntactic properties with many other verbs. Moral obligation is expressed by deber 'ought', and external obligation, weakly by haber de, and more strongly by tener que, both translatable as 'to have to'. Deber, which seems to have had a continuous history as a deontic modal, has also developed an epistemic sense which, at least in the simple paradigms, is formally marked: debe matricularse 'he must register (= is obliged to)', debe de matricularse '(I infer that) he must be registering'. An interesting parallel can be found in the Romance replacement futures (see p. 57), which began as periphrases marking weak deontic modality and have now generally developed an epistemic sense: estar a en el despacho 'she must be in the office'. All varieties of Spanish preserve a vigorous subjunctive mood (see the charts of regular and irregular verbs for the morphology). Opinion is divided, however, on whether the subjunctive should be viewed as an independently meaningful category or as a 'mere' marker of subordination. Its use in many contexts is undoubtedly determined by grammatical factors; for example, a clause dependent on querer or any other expression of volition invariably requires a subjunctive — quiero que lo hagas/*haces 'I want you to do it'. In others, the conditioning is more subtle: indicative sale 'she goes out' certainly means something different from subjunctive isaIga! 'get out!', but it has been argued that the imperative sense derives from an Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 114 SPANISH underlying volitive, '(I want you to) get out', in which the presence of the subjunctive is more readily explicable. Something similar happens in a subset of restrictive relatives, where subjunctive busco un hombre que hable árabe 'I'm looking for a man who speaks Arabic' (= any Arabic speaker) contrasts with indicative busco a un hombre que habla árabe 'I'm looking for a (particular) man who (I know) speaks Arabic'. Here it could again be argued that the alternation of indicative and subjunctive correlates with the presence or absence of 'personal a’ which establishes the referentiality of the antecedent noun. There are, however, a few instances where a genuine alternation is possible: ¿crees que vendrd? and ¿crees que venga? can both be translated as 'do you think he'll come?', but the first is neutral in implicature while the second conveys the speaker's belief that he won't. If such examples are taken as criterial, the 'grammatical marker' hypothesis cannot be sustained. Nevertheless, it remains very difficult to find a single, uniform meaning for the subjunctive, the traditional labels of 'doubt' and 'uncertainty' matching some, but by no means all, of its functions. Word Order and Cohesion Let us now turn to the combinatorial possibilities of the nominal and verbal groups we have examined separately. The minimal complete utterance, as we have seen, consists of a single verb, but a verb inflected for person and number. Most analysts assume this marking is not autonomous, but 'copied' from an underlying subject by virtue of the complex system of concord inherited from Latin. Although some exponents of concord, notably nominal case, have been eliminated during the history of Spanish, the system still marks number and gender on all modifiers within the noun phrase, and number and person (and in analytic passives gender too) between subject and verb. Concord thus guarantees the cohesion of sentential constituents and in most cases unambiguously assigns a subject to a verb. Subjects and objects that are morphologically identical can almost always be differentiated by the presence or absence of determiners and 'personal a’, or by their position relative to the verb. In Spanish simplex sentences, nominal objects and complements always follow their verb: Elena compro un coche 'Helen bought a car', el libro parecia interesante 'the book seemed interesting'. In everyday language, the VO/VC order is fixed; objects cannot precede their verbs, *Elena un coche compro. It is certainly possible to topicalise an object consisting of a definite noun phrase or a proper noun by moving it to the front of the sentence, but when this happens there is an intonation break after the topic, and an object clitic is obligatorily inserted before the verb: el coche, lo compro Elena '(as for) the car, Helen bought it\ The result is no longer a simple sentence; lo compro Elena is a complete structure in its own right. In the light of these observations and the typological statements at the Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 115 head of this section (pp. 103-4), it is puzzling that Spanish has acquired a reputation for free, or comparatively free, word order. If VO order is obligatory, and the internal structure of noun phrases largely invariable, where lies the freedom? It is due to one salient characteristic which differentiates Spanish sharply from French but is common to most of the southern Romance group, namely that subject NPs are not fixed by grammatical requirements at a particular point in the sentence. SVO, VSO and VOS orders are all found, their syntactic transparency being guaranteed in almost all cases by the operation of concord, itself corroborated by unambiguous suffixal inflections. Of course, the position of subject NPs is not random: it is heavily influenced by pragmatic and sometimes stylistic considerations. As a general rule, topics precede comments, and new information is located towards the end of the sentence. Because of the marked tendency in speech for the topic to coincide with the grammatical subject, SV order is probably the most frequent, especially if the subject consists of a single proper noun or a very short phrase. So ?compró Elena el coche would sound very odd in speech, and compró el coche Elena would tend to be reserved for contradiction or contrast: 'It was Helen (not Jane) who bought the car'. Nevertheless, VSO and VOS orders are common in more formal registers, and in all registers VS is obligatory in existential statements: viven gitanos en las cuevas 'there are gypsies living in the caves'. There is also a strong tendency for unusually long or 'heavy' subject phrases to appear on the right of the verb, han llegado (en Madrid) todos los transeuntes de la Compania X 'all passengers travelling with Company X have now arrived (in Madrid)' and VS order is the norm in many types of subordinate clause even when the subject consists of a single word: no vi lo que leia Juana 'I didn't see what Jane was reading'. The mobility of subjects is not, in itself, a serious difficulty for a typological description and does not affect most of the features implied by VO structure, but it does cause problems in assigning a 'basic' word order to Spanish: SVO, VSO and VOS are all possible candidates. For generative linguists, the mobility of the first NP has proved more problematic, since both leftward and rightward movement rules are in this case difficult to constrain, and a rightward rule would often break up the presumed inviolability of the VP constituent. We have so far avoided the term 'verb phrase', precisely because the existence in Spanish of such a sentential constituent remains undemonstrated and possibly undemonstrable. Aside from subject NPs and adjectives, the elements enjoying most syntactic freedom are adverbial clauses and phrases, and the quantifier todo 'all' which may 'float' from a subject NP to postverbal position. In conditionals, for instance, the protasis normally precedes the apodosis, but either order is acceptable. Likewise, temporal adverbial clauses tend to precede the main clause as an iconic reflection of chronology, but saliluego que hubo terminado la cena i went out as soon as he'd finished eating Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 116 SPANISH supper' is perfectly grammatical in Spanish, as in English. Simple adverbs, on the other hand, tend to be functionally located and any movement, when possible, may affect their scope relations. Thus, an adverb acting as a sentence modifier is usually placed in initial position — Desgraciadamente ... 'Unfortunately ...' — and its scope extends rightwards throughout the sentence; an adverb modifying an adjective similarly precedes and has rightward scope — muy perezoso/*perezoso muy 'very lazy'; but an adverb qualifying a verb normally follows and therefore has leftward scope — corre rdpidamente/? rdpidamente corre 'runs fast'. Negatives and Interrogative Negatives and interrogatives have virtually no freedom of word order. Spanish is a 'wh -fronting' language which obligatorily 'piedpipes' any accompanying preposition to initial position: icon quien salieron?"who did they go out with?/with whom ...?'. ¿salieron con quien? is permissible only as an incredulous echo question and *¿quien salieron con? is ungrammatical. Yes-no questions in speech often rely entirely on intonation to differentiate between interrogatives and statements. Both VS and SV orders are found, but VS cannot be assumed to be the result of syntactic inversion, since VS order is usually acceptable in the corresponding statement. VS order is normal in questions beginning with an interrogative word: ¿qué quieren ustedes? 'what would you (pI.) like?', but some Caribbean varieties regularly show an SV order, ¿qué ustedes quieren?, that would be rejected by the metropolitan standard. Sentential negation in Spanish is expressed by inserting no immediately before the verbal group (including clitics). Anything to the right of no may fall within its scope. If the intention is to negate the verb contrastively, a break in intonation or a syntactic permutation like topicalisation may be used to exempt items normally to the right of the verb from the unwanted negation. Phrasal negation is likewise achieved by a preposed no and demarcative intonation. (Occasionally no is found after quantifiers and in sentence-final position, apparently with leftward scope, as in los otros merendaron pero yo no 'the others had tea, but I didn't'; but such examples are best treated as ellipses.) Multiple negation is quite acceptable in Spanish: no dice nunca nada a nadie 'he never says anything to anyone' (literally 'he doesn't never say nothing to nobody'). For negative pronouns and adverbs, there is a choice of construction: 'nobody came' can be rendered either as no vino nadie or nadie vino, but some form of preverbal negation is obligatory — *vino nadie is ungrammatical. Since multiple negation is the norm rather than the exception, extra negatives cannot be added in order to intensify the effect. The curious alternative is to combine a preconstituent negator with a following positive modifier: sin duda alguna 'without any doubt/without the slightest doubt'. Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 117 Complementation Several aspects of complex sentence structure have by now been mentioned under different headings, notably relativisation in connection with noun phrase organisation and modality (pp. 106 and 114) and adverbial clauses in connection with word order constraints (p. 115). In some others, particularly coordination, Spanish does not differ significantly from Latin or from its Romance neighbours. In the structure of complementation, however, there have been some notable changes. The principal classical Latin patterns of accusative + infinitive, or finite subjunctive clauses in continuous indirect speech, have given way progressively to analytic constructions on the model of QUOD > que + indicative, or to infinitival complements (see pp. 65ff). In modern Spanish, verbs of saying, thinking and believing take the que complementiser followed by a finite indicative while those suggesting a stronger emotional state, such as volition, exhortation, demand or fear, require que plus a finite subjunctive. In both cases there are 'sequence of tense' constraints which limit the range of tense-forms in the subordinate clause by reference to the main verb. In both cases also, the finite clause may be substituted by an infinitive when the subjects of both clauses are coreferential. Yet this strategy is used much more frequently to avoid a subjunctive clause than an indicative. Quiero que (yo) lo haga 'I want that I should do it', though grammatical, is clumsy and improbable; quiero hacerlo I want to do it' is far more natural. On the other hand, to replace creo que me he equivocado i think I've made a mistake1 by creo haberme equivocado signals much greater formality of register. So too does the last remnant of the Latin accusative + infinitive in: la creo inteligente 'I believe her (to be) intelligent', as opposed to creo que es inteligente 'I think she's intelligent'. Despite this minor anomaly, infinitival complementation with coreferential subjects has made great strides in Spanish, to the point where the infinitive has become one of the most frequent forms of the verb. Complementation of this kind often requires a linking' preposition. Sometimes, the presence or absence of a linkage appears to correlate with some other grammatical property: for instance, all lexically reflexive verbs require a preposition, perhaps as a signal that the infinitive does not represent the direct object. Occasionally, the preposition itself continues an earlier metaphor which has now become opaque, as in abstenerse de hacer 'to abstain from doing' (= 'to hold oneself away from'). Sometimes, too, the preposition echoes the derivational structure of the verb, as in insistir en 'to insist on'. But more usually the linkage seems quite arbitrary and must be learned as an idiosyncratic feature of the verb. There seems to be no semantic reason for decidir 'to decide', intentar 'to try' and pensarmth the meaning 'to intend', all to take a bare infinitive, when decidirse and resolverse 'to decide/resolve' both require a, optar 'to opt' requires pory tratar 'to try' requires de, and pensar with the meaning 'to think about' requires en, even Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 118 SPANISH though 'to dream about' is sonar con. It is not unusual for languages to be left with arbitrary anomalies of this kind as a result of divergent historical processes, but it is very puzzling to find a language creating irregularity on this scale. So far, no satisfactory explanation has been proposed. 5 Lexis Spanish, like other world languages, preserves in its lexicon a multi-layered record of historical contacts, both direct and indirect. For ease of exposition we shall divide Spanish vocabulary into three categories: inherited, borrowed and created items. It must, however, be borne in mind that these diachronic categories are by no means always delimitable by criteria of synchronic analysis, nor recognisable to untrained native speakers. Inherited Items By 'inherited' vocabulary is meant: the stock of Latin words introduced during the main phase of colonisation and continuously in use from that date to the present, plus the few items surviving from the pre-Roman substratum. Since little is known of the indigenous Iberian languages, attribution of a term to this source is often rather tentative, when other origins have been investigated and eliminated. Such etymological puzzles include: barro 'mud', cama 'bed', the adjective gordo 'fat', manteca 'lard', paramo 'moor', pizarra 'slate' and vega 'river plain', hquierdo 'left' has congeners throughout the peninsula and in south-west France, and is cognate with Basque ezker, but may be a borrowing there too. No satisfactory etymology has been found for perro 'dog', which until the sixteenth century seems to have been a popular alternative to Latin can, now restricted to a few rural and astronomical contexts. A number of common words derive from Celtic, including: cambiar 'to change', camino 'road', camisa 'shirt', cerveza 'beer', and carro — 'cart' in Spain, though in much of Latin America it has come to mean 'car', by contamination from English (see p. 241). These and similar words, however, are widely distributed in Romance and may therefore be borrowings into Latin from Gaulish. Direct survivals from the Celtic spoken in north and central Iberia probably include: Sp., Port, alamo 'poplar', Sp., Port, gancho 'hook', Sp. engorar, Port, gorar 'to addle', Sp. serna, Port, seára 'sown field', with berro 'watercress' and légamo 'slime' apparently exclusive to Spanish. Borrowings A significant number of Germanic words remain in regular use, nearly all of them widespread in Western Romance; they were probably diffused via Latin before the main period of invasions, as a result of Roman contact with 'confederate' Germanic peoples. They include military terminology — guerra 'war', guardia 'guard', tregua 'truce', espuela 'spur', estribo 'stirrup', Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 119 yelmo 'helmet' — but also some everyday words, like ropa 'clothing', falda 'skirt', jabón 'soap', ganso 'goose', ganar 'to win', together with a set of common adjectives, fresco 'fresh/cool', rico 'rich', bianco 'white', and gris 'grey'. Some other words of ultimately Germanic origin were borrowed much later from French, Occitan and Catalan, including: orgullo 'pride', galardon 'reward' and guante 'glove'. Direct contact between Germanic and Hispano-Romance speakers seems not to have been extensive. The early waves of invaders left few permanent settlements and the later Visigoths, who ruled substantial areas of Hispania for over two centuries, had long been in contact with Romans first as confederates then as founders of the kingdom of Toulouse, and were almost certainly bilingual, if not monolingual, Romance speakers at the time of their arrival in the peninsula (see p. 2). It is in place and personal names that Germanic elements have most vividly survived. Andalucia owes its name (after a little Arabic readjustment) to the Vandals who briefly settled there; numerous villages are called Suevos (after Swabians) and Godos (after Goths) or some still-recognisable variant like Villagodos. Common personal names include: Alfonso, Alonso, Alvaro, Elvira, Fernando, Gonzalo, Ramon and Rodrigo. It is likely that the Visigoths were responsible for the increase in use of the patronymic suffix -ez (probably calqued from the Latin genitive -IC1) in, for instance: Fernandez, Gomez, Gonzalez, Hernandez, Lopez, Martinez, and so on. Approaching four thousand words can be traced to Arabic, almost all of them nouns and a high proportion beginning with a- or al~, representing the agglutination of the Arabic definite article. An important group relates to horticulture and water management: acequia 'irrigation channel', noria 'water wheel', aljibe 'cistern', aceite 'olive oil', alcachofa 'artichoke', algodon 'cotton', arroz 'rice', azafrdn 'saffron', azúcar 'sugar', berenjena 'aubergine', naranja 'orange', zanahoria 'carrot'. Others concern civil administration: aduana 'customs', alcaide 'governor/gaoler', alcalde 'mayor', alguacil 'constable'; and still others have been assimilated, via Spanish, into international scientific vocabulary: alcohol, algebra, cifra 'figure/cipher', cenit 'zenith', nadir, and so on. In southern Spain, where the period of contact was longest and most intense (see pp. 6-7), some of the features most often cited as 'typical' are designated by Arabic words: azahar 'orange blossom', azotea 'flat roof, azucena 'lily', and azulejo 'ceramic tile' (so called because the basic colour was a deep blue — azul). In categories other than nouns, Arabic has given the adjectives baldio 'fallow' and mezquino 'mean', the verbs achacar'Xo accuse' and halagar 'to flatter' (all of them well adapted to Romance grammatical patterns), the preposition hasta 'up to', and the exclamative ojald 'would that ...' (literally: 'may Allah grant ...'). An interesting disputed case is that of usted(es), the polite form of address (see pp. 96-7), often explained as a contraction of vuestra merced(= approximately 'Your Grace'; compare Port. Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 120 SPANISH você, Cat. vostè), but now thought to have been blended with Arabic /ustad/ Iord, master'. During the colonial era, from the sixteenth to mid-nineteenth centuries, words were assimilated from Amerindian languages, those denoting new concepts often going on to become internationalisms: cacao, chicle, chocolate, coyote, tomate, from Nahuatl; maiz, patata, tobaco, from Arawak; alpaca, cóndor, llama, from Quechua; anand(s) 'pineapple', petunia, tapioca, tucdn, from Guarani. In general, Castilian has retained Amerindian terms only for new concepts; a rare exception is Nahuatl tiza 'chalk' (as a writing material) which has restricted Romance greda < CRETAM to the rock itself. As is to be expected, Latin American varieties have borrowed more extensively from the indigenous languages with which they continue to be in contact. In much of Mexico, for instance, Nahuatl tianguis is the ordinary word for 'market', and zacate and zopilote may be preferred to designate 'grass' and 'vulture' respectively. Likewise in common usage in Peru are some Quechuan words (charaque 'cured meat', chuho 'potato flour', poroto 'bean', soroche 'mountain sickness') and in Paraguay others from Guarani (cobaya 'guinea pig', tapera 'ruins, debris', yacarelyaguare 'alligator'). Since the Renaissance, Spanish has enriched its lexicon, mainly in the formal and technical registers, by extensive borrowing from Latin and other Romance languages; and to the present day, Latin and Greek continue to be favoured sources for the creation of new terminology. (We shall examine below some linguistic consequences of widespread borrowing from cognate languages.) Borrowing from English, relatively minor before the nineteenth century, has accelerated rapidly in the twentieth and affects all varieties, most particularly those in the Caribbean and Central America in regular contact with the USA. Although the recent influx of words (and cultural artifacts) from this source has been considerable, the proportion of non-specialist words permanently incorporated into Spanish is likely to be much lower than purists fear. Degrees of Assimilation The degree to which borrowings are intuitively recognised as intrusive depends on such factors as their adjustment to native phonological and orthographic patterns, their date of acquisition, and the level of education of the person making the judgement. Spanish school children are sometimes taught to think of j and z as 'the Arabic consonants' especially when both occur in the same word; while there is some truth in the generalisation, it also makes false predictions — so ajimez 'window arch' and ajedrez 'chess' are correctly identified as Arabic but juez 'judge', from IUDICEM, is wholly Romance. One possible measure of assimilation is participation in regular morphological processes. Notice that the vast majority of borrowings are nouns and as such require relatively little morphological Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 121 adjustment; most participate readily in plural inflection, but a recent acquisition like jersey can pluralise either regularly as jerseyes or as jerseys, presumably in imitation of English. Derivation is a particularly effective means of integration: the Germanic borrowings quoted above, guerra and orgullo regularly form adjectives guerreroand orgulloso; the Arabic halagar forms halagueno; and in much of Central America a 'peasant farmer' is called milpero, an agentive derivation from the indigenous word milpa 'maize field'. The gradual blurring of distinctions between inherited and borrowed categories, often abetted by lexical creativity, confronts descriptive linguists with an intriguing problem in the specific case of late borrowings from Latin, many of which are related to words that have had a continuous history in the language. This relationship, however, is no longer immediately apparent because the inherited items have generally undergone more extensive phonological modification than the late-comers, which have tended to be admitted in a hispanised pronunciation of the original spelling. Some idea of the scale of borrowing of 'learned' vocabulary from Latin can be gained from the list given here, which shows only a subset of the common verbs and nouns that underwent the phonological change f- > h> zero, a change completely bypassed by their associated high register adjectives. The question is therefore whether the relationship is psycholinguistically real and, if so, whether it should be recreated by productive rules in synchronic morphophonemics. While attempts to relate 'doublets' of this kind are particularly associated with the abstract phase of generative phonology and are nowadays less often attempted or even deemed desirable, we should note that any across-the-board phonological solution, whether concrete or abstract, is liable to run foul of mixed derivational sets: hwno 'smoke' has as its regular derivatives humoso 'smoky' and ahwnar 'to preserve food by smoking', but it is also manifestly related to fumar'to smoke' (of fires or of people). Popular and Learned Word Pairs hicrro 'iron' hijo 'son' hado 'fate' hambrc 'hunger' harina 'flour' hastio 'distaste' férrico filial fatal famélico farináceo fastidioso heder 'to stink' hembra 'female' hongo 'mushroom' hormiga 'ant' huir 'to flee' hurto 'theft' fétido femenino fungoso fórmico fugaz furtivo A further example of overlapping categories and progressive reanalysis is provided by Spanish negatives (see p. 116 for syntactic aspects of negation). Most derive ultimately, but not directly, from Latin NE, once the universal negator, but by the classical period more familiar as a prefix creating a whole family of compound negatives, including NON < NE + OINOM 'not Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 122 SPANISH one, not a single'. This resulted in a series of synchronic alternations, all of which have been disrupted by subsequent lexical replacements. So, USQUAM 'anywhere'/NUSQUAM 'nowhere' were eliminated before the earliest Spanish texts in favour of periphrases meaning 'in any place/in no place'; UMOUAM 'ever' was likewise rejected in favour of a periphrasis meaning 'at no time', despite the continuing vigour of NUMOUAM 'never' as nunca; and ULLUM 'any' gave way to ALICUNUM > alguno, despite the survival into Old Spanish of NULLUM 'no, not any' as nul (modern nulol nulidad are much later high-register borrowings). One might thus expect the derivational transparency of Latin negatives to have been edited out of Spanish; but this is not entirely the case. In Old Spanish, the basic negators non and ni 'neither, nor' < NEC (< NE + QUE), together with nunca and nul, seem to have been perceived as a synchronic set. One indication is that non, whose final -n persisted until the early fifteenth century, apparently exerted an analogical pull on ni, which is frequently written nin during the same period. Likewise neguno < NEC + UNUM, the Romance rival to nul, is attested from the earliest texts with the alternative spellings niguno and ninguno. Also Romance creations were the pronouns nada 'nothing' and nadi 'nobody', both variants of the past participle nado of the verb nacer 'to be born', used colloquially as an intensifier ('not a born man would do it', 'we saw not a born thing') but having no etymological connection with negation. The success of nada and nadi is due to a combination of factors: their reinterpretation as negatives was provoked by frequent collocation, later colligation, with the true negator no, a development paralleled in French (see p. 237) but boosted in Spanish by the quite fortuitous initial /n-/. Their consolidation was assured when nado was displaced as the past participle of nacer by the analogical creation of a regular form nacido, which effectively severed the lexical and metaphorical links between the verb and the new negative pronouns. The modern Spanish negatives are thus the result of inheritance, analogy and sheer accident. Even so, no, ni, nada, nunca, ninguno and ninguna parte 'nowhere' (the analytic replacement for NUSQUAM) give an appearance of systemic cohesion. The pattern in turn has been bolstered by lexical survivals, borrowings and derivations. Negar 'to deny' < NEGARE is attested in Old Spanish, with derivatives negativo and negacion following in the fifteenth century, and negable much more recently. Neutro 'neuter' < NEUTRUM (literally 'neither one nor the other') is not attested before the fifteenth century, but has now acquired: neutral, neuttalidad, neutralizar, and neutron. Alongside the obviously learned borrowings nefando/ nefasto 'nefarious' (literally 'not to be spoken of), nihilismo and nimio 'excessive', we find the more popular coinings nonada/naderia 'trifle, mere nothing'. The consequence is that words of very different dates and provenance now conspire to associate initial /n-/ with an obvious semantic component of negativity, to the point where it is most unlikely that native Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 123 speakers would fail intuitively to recognise the coincidence. Whether /n-/ can be considered to be a synchronic morpheme is more doubtful, since it is no longer productive (unlike, for instance, the negative prefixes in- and des-) and its segmentation would create a large number of unanalysable roots with unique or near-unique distributions. It seems best, therefore, to treat /n-/ as a submorphemic marker of negation. Creativity and Rates of Replacement The proportions of inherited, borrowed and created words within the total word stock differ appreciably from their relative frequencies in continuous discourse. In the total stock, for instance, the words directly inherited from Latin constitute a small, if important, minority; they are proportionately much more frequent in discourse. This is especially true for functors (grammatical items such as determiners, pronouns, prepositions and conjunctions), which represent a tiny proportion of the total, but can nevertheless account for over two thirds of the words in running text. A recent study of the 5,000 most frequent words found that they were composed of some 24 per cent inherited items, 35 per cent created and 41 per cent borrowed (with both the latter categories in turn heavily reliant on Romance roots), but that the 24 per cent of inherited words accounted for 81 per cent of all occurrences. Functors make a major contribution to this high total, all but one — the Arabic preposition hasta 'until’ — being derived from Latin roots. Figure 3.4: Lexical Relationships % cognacy Castilian to Portuguese Castilian to Catalan Castilian to Gascon Shared cognacy (modern) Castilian to Classical Latin Shared cognacy (achronic) Strict cognates Root cognates 0 10 20 30 40 50 60 70 80 90 100 Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 124 SPANISH This impression of relative conservatism in the lexicon (or, alternatively, of a low replacement rate), is corroborated by the standard lexicostatistical test of 100 'core' concepts, which in turn provides a yardstick for comparing Spanish with neighbouring varieties of Romance. Figure 3.4 shows the relative cognacy rates as established by this method. The percentage of shared items for each pair of languages varies depending on the inclusion or otherwise of root cognates. For example, Port, frio, Sp. frio, Cat. fred, Gasc. hred, Fr. froid, It. freddo, Rum. frig 'cold' are all fully cognate with Latin FRIGIDUM; whereas Port, cabeca, Sp. cabeza 'head' are fully cognate with each other, but only root cognates of Cat., Gasc. cap, and CLat. CAPUT. By the looser criterion, modern Spanish still shares 75 per cent of its core vocabulary with Classical Latin, a figure which falls only to 64 per cent if the stricter criterion is adopted. This already high retention rate could, however, be increased if instead of comparing Spanish with Classical Latin, we compared it with the presumed lexicon of Proto-Romance. In Figure 3.4, the discrepancy of eight items between the synchronic and achronic cognacy rates, is almost certainly due to shared innovations of early Romance rather than to subsequent borrowing. Five of the eight result from semantic shifts or lexical replacements using Latin roots, for example FOCUM, originally 'hearth', displaces classical IGNEM 'fire' to give Port, fogo, Sp. fuego, Cat. foe, Gasc. huec; and BUCCAM, originally 'cheek', displaces OREM 'mouth' to give Port., Sp., Cat., Gasc. boca. Two involve derivation without lexical change: Port, orelha, Sp. oreja, Cat. orella, Gasc. aurelha 'ear', all derive from AURICULAM rather than classical AUREM. Only one is an external borrowing: the Germanic word BLANK 'white' gives Port, branco, Sp. bianco, Cat./Gasc. blanc, though reflexes of classical ALBUM survive in literary registers (Sp. alba 'dawn') and in numerous derivatives, both popular and technical (Sp. albero 'tea/dish towel', albuminoide 'albuminoid'). In terms of synchronic relationship, the lexical overlap of Spanish and Portuguese is very striking, at 89 per cent strict cognacy and 92 per cent root cognacy. Spanish is progressively less well related to Catalan (72 per cent strict cognacy), Gascon (64 per cent) and standard French (61 per cent), along a scale which obviously correlates with geographical distance. In the few items of core vocabulary where Spanish and Portuguese differ, Spanish is usually the innovating member of the pair, though not always in a way that matches its easterly neighbours. So, Sp. perro 'dog' (origin uncertain; see above) contrasts with Port, cao and Gasc. can < CANEM, but also with Cat. gos; while Port, joelho, Cat. genoll, Gasc. jolh < GENUCULUM 'knee' all contrast with Sp. rodilla, originally a metaphor 'little wheel, pivot' (compare Fr. rotule 'kneecap') which took over when OSp. hinojo became homophonous with hinojo < FENUCULUM 'fennel'. A more frequent pattern, however, finds both Spanish and Portuguese behaving with the conservatism typical of peripheral areas, in face of cen- Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 125 tral innovations which peter out in Gallo-Romance or Catalan. Examples are: Port, pássaro, Sp. pájaro < PASSAREM 'bird' versus Cat. ocell, Gasc. ausèth < AVICELLUM; Port., Sp. comer < COM+EDERE 'to eat' versus Cat. menjar, Gasc. minjar < MANDUCARE; Port, areia, Sp. arena < ARENAM 'sand' versus Cat. sorra, Gasc. sable; Port, pequeno, Sp. pequeño 'small' verus Cat., Gasc. petit (classical PARVUM does not survive); Port., Sp. rabo < RAPUM 'tail' verus Cat. cua, Gasc. coda < CAUDAM; and Port, mulher, Sp. mujer < MUUEREM 'woman' versus Cat. dona, Gasc. hemna. In rarer instances, Portuguese and Spanish have innovated eccentrically: Port., Sp. garra 'claw', probably an Arabic-Romance blend, contrasts with Cat., Gasc. urpa; and Port, amarelo, Sp. amarillo 'yellow', derivatives of AMARUM 'bitter', contrast both with Cat. groc < CROCEUM 'saffron coloured' and with Gasc, Fr. jaune < GALBINUM 'yellowish green'. Creative Mechanisms In previous stages of the language, the gender markers -o/-a (see pp. 93-4) have been highly productive. In the kinship system, for example, mechanisms as diverse as ellipsis, analogy, back formation and borrowing have reduced the irregularities and suppletions of Latin, to the benefit of the -o/ -a alternation. Adjectival ellipsis is responsible for Sp. hermano/hermana < (FRATREM) GERMANUM/(SOROREM) GERMANAM 'real, full (brother/ sister)' (i.e. siblings with both parents in common); for cuñado/cuñada 'brother/sister-in-law' < (FRATREM) COGNATUM/(SOROREM) COGNATAM (originally 'half-brother/sister'); and for primo/prima 'cousin' < PRIMUM (SOBRINUM»/PRIMAM (SOBR1NAM) (literally 'first (cousin)'). These ellipses in turn eliminated the FRATER : SOROR suppletion (though the roots survive in religious contexts as fraile 'monk' and Sor Merced 'Sister Mercy'); freed sobrino/sobrina to take over the meanings 'nephew/niece'; but required the use of a derivative PRIMARIUM > primero 'first' to form an ordinal distinct from the new kinship term. Analogy is responsible for suegra 'mother-in-law' < SOCRAM, reformed from the apparently irregular fourth declension SOCRUM to match suegro 'father-in-law' < SOCERUM. Nieto 'grandson' appears to be a back formation from nieta 'granddaughter' < NEPTAM, itself an analogical readjustment of classical NEPTEM. Tio/tia 'uncle/aunt' are borrowings from Greek. Against this generalisation of the -of -a pattern, we should note the persistence of suppletion in the most common kinship pair: padre 'father' < PATREM : madre 'mother' < MATREM. Among animal names, suppletion has likewise persisted in tow 'bull' < TAURUM: vaca 'cow' < VACCAM; and one has been introduced by the adoption of caballo 'horse' < CABALLUM (originally 'nag') to replace classical EQUUM, even though EOUAM 'mare' survives as yegua. An aspect of the -o/ -a alternation that has proved highly contentious is the treatment of a small number of Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 126 SPANISH masculine/feminine pairs like barcolbarca 'ship' and cesto/cesta 'basket'. In these, the feminine is usually the older form, and the denotational differences relate to size, shape and function; but the correlation of these features with gender is, at best, tentative. Overall, the -o/-a alternation, though extremely frequent in Spanish, has long passed its peak as a creative mechanism and seems now to be on the verge of non-productivity. Many recent neologisms are neutral in gender — like astronauta 'astronaut', izquierdista, 'leftist', racista 'racist' — and rely on the determiner to mark the sex of the referent, if need be. Since all neologisms of this kind take plural marking, the trend does not seem to signal a more general move towards morphological invariability. Created Items: Derivation and Composition Throughout the history of Spanish, derivational processes have played a major role in extending the lexicon and increasing the range of 'motivated' vocabulary. (Conversely, the widespread borrowing of partial synonyms of higher register that we examined above, may well have increased lexical opacity.) Prefixation, very common in Latin, has left numerous traces in inherited vocabulary, though it is not always clear that segmentation is synchronically valid: concebir 'to conceive', percibir 'to perceive' and recibir 'to receive' embody prefixes which are still productive (very much so in the case of re-), but which, if segmented here, neither have their modern meanings nor leave a coherent semantic value for the root. Recently, prefixation has regained considerable momentum. Super-/hiper- can now be prefixed to almost anything as superlatives (supermercado 'supermarket', hiperraquitico 'extremely thin/weak'); smallness is indicated by mini{minibus, minitren 'minitrain, shuttle'); forward position by ante- (anteojos 'goggles') or pre- (preescolar 'pre-school'); opposition or protection by anti(antinuclear, antirresbaladizo 'antiskid', faro antiniebla 'fog lamp'); and reflexivity by auto- (autoservicio 'self-service', autodestruccion 'selfdestruction', autocargante 'self-charging'). Infixation is a rarer process, and one open to competing interpretations, as is so-called 'parasynthesis'. Both are illustrated by enfurecerse 'to get angry', transparently derived from furia, but variously argued to result from the conscious use of an infix (-SC-; see p. 101) or of a reanalysed inchoative suffix (-ecer), coupled simultaneously or later with a prefix and reflexivisation. A similar difficulty is posed by verbs like lloriquear 'to whimper', which may appear to have a pejorative infix when compared with a base form like llorar 'to weep', but are probably better analysed as having one of a range of affective suffixes which behave in morphologically similar, but not identical ways. So, traquetear 'to rattle' along with its noun traqueteo is related to the onomatopoeic trac 'clang', but has no base verb on which infixation could operate. Similarly, lloviznar 'to drizzle' is clearly related to the basic llover 'to rain', but the change in conjugation membership implies the use of a Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 127 ready-made suffix, rather than a two-stage process of infixation. Suffixation has always been particularly vigorous in Spanish, as a method of both changing morphological class and adding expressive nuances. Among function changers we should note: adjectivals -ero, -oso, -able/Able(inoxidable 'stainless', but also presidenciable 'suitable to be president'); the abstract nominals -anciaf-encia, -acion/icion, -ez, -miento and -aje (borrowed from French and cognate with the rather rare negative -azgo; compare andamiaje 'scaffolding' with noviazgo 'courtship'); the vogue nominal -ota (pasota 'drop-out', drogota 'junkie'); and the agentive -dor, which in the feminine and after the ellipsis of mdquina 'machine' has coined labels for a whole gamut of labour-saving devices {afeitadora 'shaver', embotelladora 'bottling machine', licuadora 'liquidiser', secadora 'spin dryer', among many others). Spanish also possesses a bewildering array of so-called 'diminutives' and 'augmentatives', but these seem always to have carried affective connotations rather than strict denotation of size. Still highly productive are the diminutives -ito, -ino (usually implying affection) and -Mo (often, though not always, pejorative). The augmentatives -on and -azo generally indicate some kind of excess rather than mere size: hombron 'hulk of a man', solterona 'old maid'; but the excess can also connote approval, as in colloquial molon 'great, fantastic', from molar 'to go well, really swing'. Some idea of the mixed flavours of these suffixes can be gauged from the series: casa 'house', casita 'cottage', caseta 'weekend cottage', casMa 'hut', casuca 'shack' (both affectionate and pejorative depending on context), casucha 'hovel', cason 'rambling country house', caseón 'dilapidated old pile'. Nor are such suffixes confined to nominal stems; adjectives and adverbs are quite frequent hosts — gordo 'fat' : gordito 'chubby', ahora 'now' : ahorita mismo 'this very second', despacio 'slowly' : !jdespacito! 'gently does it!'. By comparison with the exuberance of derivational mechanisms in Spanish, compounding is often unjustly represented as a minor process. In fact, compounding has been used as a source of neologisms at all stages of the language, and recently there has been a notable upsurge in the process. Long-established two-term patterns are illustrated in: agridulce 'bittersweet', bienvenida 'welcome', caradura 'cheek, insolence' (literally 'hard face'), dondequiera 'wherever', madreselva 'honeysuckle' and mediodia 'midday'. Phrasal compounds include: correveidile 'gossip' (literally 'rungo-and-tell-him/her'), hazmerreir 'jester, laughing stock' (literally 'makeme-laugh') and sabelotodo 'know-all'. One very productive pattern creates a compound adjective from a part of the body plus its attribute: pelilargo 'long-haired', cejijunto 'beetle-browed', labihendido 'hare-lipped', barbirrojo 'red-bearded', piernicorto 'short-legged'. But by far the most common mechanism for recent neologism is the compounding of verb -h object, where the object is often plural, of either gender, but the resulting noun is masculine singular. Well entrenched examples are: abrelatas 'tin-opener', Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 128 SPANISH cascanueces ‘nut-cracker', engañabobos 'booby-trap', espantapájaros 'scarecrow', sacacorchos 'cork-screw'; together with humorous pejoratives like sacamuelas 'dentist, tooth-puller'. We end this section on a note of optimism. Although purist hackles have been raised by the recent influx of anglicisms (as in France, see p. 243), the productive patterns of the language remain resolutely Romance. The best evidence is that new concepts and artifacts which might easily have attracted a foreign label are so often named from indigenous roots, whether by derivation or compounding. Urbanizatión 'housing development', currently to be seen on builders' placards all over Spain, is made up of impeccably classical roots. Calientaplatos 'plate-warmer', lavaplatos 'dishwasher', limpiaparabrisas 'windscreen-wiper', and even, alas, cartabomba ietter-bomb' use only indigenous material. Through developments of this kind, Spanish is becoming more, not less, Romance in its structure. 6 Conclusion Comparing the two principal norms of Spanish, we have noted a number of differences in phonology, morphosyntax and lexis, but have nevertheless argued that the mainstream varieties can be grouped together as a diasystem, a shared underlying structure with relatively minor surface divergences. How has this come about? And where should the diasystemic limits be set? If membership of hispanidad is determined by mutual intelligibility, we are obliged to exclude the Creoles of Colombia and the Far East which, though often loosely described as 'Spanish Creoles', appear on closer scrutiny to have autonomous grammatical systems (for further discussion see Chapters 1 and 12). More problematic are the 'Hispanic' varieties of the United States which range on a continuum between lightly dialectal puertorriqueño and the basilectal form of chicano, which has undergone some of the morphological modification usually associated with creolisation and has assimilated numerous caiques of American English lexical and idiomatic structures. These internal characteristics, together with the frequent code-switching between Spanish and English common to all Hispanic variants in the USA, can render chicano totally impenetrable to monolingual Spanish speakers. On the other hand, there is usually mutual comprehensibility between standard and Judeo-Spanish speakers, perhaps with some difficulty and depending — because of the large numbers of lexical borrowings from adstrate languages — on the topic of discourse. But djudeo-espanyol in present-day Israel is a cover term for a continuum of variants ranging from the standard seseante Spanish (see p. 82) delivered with a slight regional accent which is used in broadcasts by Boz de Israel radio, to the unmistakably different version chosen for the associated quarterly magazine Aki Yerushalayim which has its own morphosyntax, Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 SPANISH 129 vocabulary and highly distinctive orthography (see p. 92). Mutual comprehensibility here is less a matter of intralinguistic features than of cultural choice. The maintenance of a shared orthography and written standard (for which normative organisations such as the Real Academia de la Lengua and its corresponding Academies in Latin America must be given some credit) has probably been the crucial factor in the survival of hispanidad and in preventing a repetition of the linguistic fragmentation that followed the dissolution of the Roman Empire (see pp. 2-5). In the colonial era, the political, religious and administrative bonds linking Spain to its Latin American territories were not designed (any more than their Roman predecessors had been) to keep the colonies in close contact with one another, and there is evidence of quite rapid divergence as vocabulary, grammatical caiques and phonetic traits were assimilated from Amerindian languages. Fragmentation could easily have gone hand in hand with nineteenthcentury independence movements. But cultural and linguistic links with Spain survived political independence, and those same factors which in Europe caused the decline of rural dialects — improved communications and the spread of literacy — not only maintained the essential unity of the written language, but actually brought about convergence in formal and technical registers. Today, an Argentinian novelist can unambiguously evoke an Argentinian setting through dialogue and to a lesser extent narrative; but it is often impossible to tell, on linguistic grounds alone, whether a newspaper leader, legislative text or set of instructions for some domestic appliance, originated in Spain or in any of the nineteen Latin American countries where Spanish is the national language. Spanish is now firmly enshrined in the Constitutions of Spain and most countries of Latin America, and is viewed by the respective governments as a symbol of national unity. It is also, with Portuguese, one of only two Romance languages to be increasing rapidly its numbers of speakers; on those grounds alone its future seems assured. But in the process of expansion from minor dialect to major world language, Spanish has become a little more like some of the varieties it once rivalled. Bibliography Amastae, J., and L. Elias-Olivares (eds.) (1982) Spanish in the United States. Cambridge University Press, Cambridge. Canfield, D.L. (1981) Spanish Pronunciation in the Americas 2nd edn. Chicago University Press, Chicago. Cano Aguilar, R. (1981) Estructuras sintácticas transitivas en el espanol actual. Gredos, Madrid. Contreras, H. (1976) A Theory of Word Order with Special Reference to Spanish. North-Holland, Amsterdam. Corominas, J. and J.A. Pascual (1980-6) Diccionario critico etimológico castellano Downloaded By: 192.168.0.15 At: 12:32 12 Jun 2017; For: 9780203426531, chapter3, 10.4324/9780203426531.ch3 130 SPANISH e hispánico 6 vols. Gredos, Madrid. Cressey, W.W. (1978) Spanish Phonology and Morphology: A Generative View. Georgetown University Press, Washington DC. Fant, L. (1984) Estructura informativa en español. Estudio sintáctico y entonativo. Uppsala University Press, Uppsala. Fontanella de Weinberg, M.B. (1976) La lengua española fuera de España. América, Canarias, Filipinas, Judeo-españoL Paidós, Buenos Aires. Galmes de Fuentes, A. (1983) Dialectologia mozdrabe. Gredos, Madrid. Garcia, E.C. (1975) The Role of Theory in Linguistic Analysis. The Spanish Pronoun System. North-Holland, Amsterdam. Harris, J.W. (1983) Syllable Structure and Stress in Spanish. MIT Press, Cambridge, Mass. Haverkate, H. (1979) Impositive Sentences in Spanish. North-Holland, Amsterdam. Lapesa, R. (1980) Historia de la lengua espahola 8th edn. Gredos, Madrid. Lloyd, P.M. (1968) Verb-Complement Compounds in Spanish. Niemeyer, Tubingen. Lorenzo, E. (1971) El español de hoy, lengua en ebullicón 2nd edn. Gredos, Madrid. Luján, M. (1980) Sintaxis y semántica del adjetivo. Ediciones Cátedra, Madrid. Macpherson, I.R. (1975) Spanish Phonology: Descriptive and Historical. Manchester University Press, Manchester. Malkiel, Y. (1970) Patterns of Derivational Affixation in the Cabraniego Dialect of East-Central Asturian. University of California Press, Berkeley. (1972) Linguistics and Philology in Spanish America. Mouton, The Hague. Martin Zorraquino, M.A. (1979) Las construcciones pronominales en espñol. Gredos, Madrid. Menéndez Pidal, R. (1972) Origenes del espahol7th edn. Espasa-Calpe, Madrid. Moliner, M. (1982) Diccionario de uso del espahol 2 vols. Gredos, Madrid. de los Mozos, S. (1984) La norma castellana del espahol. Ámbito Ediciones, Valladolid. Patterson, W.T. (1982) The Genealogical Structure of Spanish: A Correlation of Basic Word Properties. Georgetown University Press, Washington DC. Quilis, A. (1981) Fonética acústica de la lengua española. Gredos, Madrid. Real Academia Española (1973) Esbozo de una nueva gramática de la lengua española. Espasa-Calpe, Madrid. Resnick, M.C. (1975) Phonological Variants and Dialect Identification in Latin American Spanish. Mouton, The Hague. Rivero, M.-L. (1977) Estudios de gramdtica generativa del español Ediciones Cátedra, Madrid Sala, M. (1976) Le Judéo-espagnol. Mouton, The Hague. (1982) El español de America: 1 Lexico. Instituto Caro y Cuervo, Bogotá. Seco, M. (1982) Diccionario de dudas de la lengua espahola 8th edn. Aguilar, Madrid. Suñer, M. (1982) Syntax and Semantics of Spanish Presentational Sentence-Types. Georgetown University Press, Washington DC. Wright, R. (1982) Late Latin and Early Romance in Spain and Carolingian France. Francis Cairns, Liverpool. Zamora Vicente, A. (1967) Dialectologia espahola 2nd edn. Gredos, Madrid.